EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Risk-aware digital twin maintenance of fiber-optic environmental monitoring infrastructure: a reproducible simulation study

Ying Zhang, Bo Xu

Frontiers in Environmental Science · 17.11.2026 · 14

DOI: 10.3389/fenvs.2026.2018811 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Environmental Science, online erschienen am 17.11.2026. Verfasst von Ying Zhang und 1 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Digital twins can support maintenance of environmental monitoring infrastructure, but dependable decisions require consistent treatment of uncertain condition, usable sensing service, and intervention timing.

Methods

In this study, we develop a risk-aware maintenance framework for a synthetic fiber-optic monitoring network. An assumed-density state estimator keeps an unbounded Gaussian working state separate from bounded physical condition. Exact moments of a thresholded service map support a one-step, moment-based coverage screen, while paired counterfactual rollouts estimate the incremental value of preventive actions using the same service and cost model as physical evaluation. The implemented heuristic combines coverage restoration with positive-value action selection; it is not assigned a global optimality guarantee. A fully executable benchmark contains 120 logical maintenance units, four hypothetical hazard contexts, and a 180-month horizon.

Results

Across 80 paired evaluation seeds, the proposed policy maintains the nominal coverage floor at lower scenario cost than reactive maintenance and an independently calibrated condition threshold. This comparison is conditional on synthetic deterioration, maintenance-cost, and consequence assumptions; it is not an estimate of savings for an operating network. However, cheaper policies without a service-floor constraint provide less service, and the calibrated condition policy provides higher mean coverage. Uncertainty removal and unmodeled process dependence cause coverage shortfalls despite passing internal screens. Additional valuation and stronger-dependence checks examine the limits of the nominal comparison. Exhaustive small-network checks also identify two cases in which greedy selection misses an attainable screening condition.

Discussion

These results demonstrate a reproducible decision architecture and explicit cost–service trade offs, not field-validated early-warning performance or transferable economic savings.

Zitieren

APA

Zhang, Y., & Xu, B. (2026). Risk-aware digital twin maintenance of fiber-optic environmental monitoring infrastructure: a reproducible simulation study. Frontiers in Environmental Science, 14. https://doi.org/10.3389/fenvs.2026.2018811

Vancouver

Zhang Y, Xu B. Risk-aware digital twin maintenance of fiber-optic environmental monitoring infrastructure: a reproducible simulation study. Frontiers in Environmental Science. 2026;14. doi:10.3389/fenvs.2026.2018811

BibTeX
@article{Zhang2026,
  author = {Zhang, Ying and Xu, Bo},
  title = {Risk-aware digital twin maintenance of fiber-optic environmental monitoring infrastructure: a reproducible simulation study},
  journal = {Frontiers in Environmental Science},
  year = {2026},
  volume = {14},
  doi = {10.3389/fenvs.2026.2018811}
}
Kostenlos starten →