Partager

Document IIF

Digital predictive maintenance solution based on OEM physic-based models and AI/ML techniques for LNG plants.

Numéro : 63

Auteurs : PORCIANI D., LAURIOLA M.

Résumé

Rotating Equipment OEMs in the oil & gas and energy industries already leverage digital monitoring tools and applications for the fleet support for customers and operators.
Such applications are usually composed by 3 modules that ensure full monitoring and prediction of assets performance degradation, in detail:
a) Anomaly detection
b) Failure mode and effect analysis with associated severity assessment
c) Maintenance task optimization
These modules are predominantly based on OEM know-how covered by design-based analytics and supported by case management systems, with the support of certain hybrid techniques for powering up their response performance with the utilization of Artificial Intelligence and Machine Learning methods1. Such capabilities enable condition based maintenance approach that allow operators to optimize maintenance costs and have visibility on unplanned event risk.
In the LNG space, with such approaches and in conjunction with certain contractual service agreements’ models, OEMs can guarantee performances with equipment reliability above 99%. The current trend in the LNG industry, progressively adopted by other critical segments such as upstream and fertilizers, is to expand such approach from critical equipment to a plant-wide model, where the same high-level data quality and performance model run by single OEMs can be expanded to other OEMs, process equipment and Balance of Plant (“BoP”) for having a one-stop solution with a system-of-systems strategy at project level, defining a Digital Predictive Maintenance Solution for LNG assets.

Documents disponibles

Format PDF

Pages : 10 p.

Disponible

Gratuit

Télécharger

Fichier PDF des diapositives de la présentation

Pages : 14 p.

Disponible

Gratuit

Télécharger

Détails

  • Titre original : Digital predictive maintenance solution based on OEM physic-based models and AI/ML techniques for LNG plants.
  • Identifiant de la fiche : 30034676
  • Langues : Anglais
  • Sujet : Technologie
  • Source : 21st International Conference & Exhibition on Liquefied Natural Gas (LNG2026)
  • Date d'édition : 05/02/2026

Liens


Voir d'autres communications du même compte rendu (66)
Voir le compte rendu de la conférence

Indexation