ARCHITECTURAL FRAMEWORK AND ADVANCED METHODS FOR UNCERTAINTY QUANTIFICATION IN INTELLIGENT DECISION SUPPORT SYSTEMS OF COMPUTER-INTEGRATED SYSTEMS FOR OIL AND GAS PRODUCTION
Measuring Equipment and Metrology · 2026 · 87 (2) · 43-52
DOI: 10.23939/istcmtm2026.02.043 ↗Kurz gefasst
Fachartikel in der Zeitschrift Measuring Equipment and Metrology, 2026. Verfasst von Dmytro Kytsmen. Die Arbeit nennt 45 Literaturangaben.
Abstract (Originalsprache)
This paper presents a generalized structural framework of IDSS_UQ together with a comprehensive analysis of advanced uncertainty quantification (UQ) methods for intelligent decision support systems (IDSS) integrated into computer-integrated systems (CIS) for oil and gas production. Represented study analyzes a broad methodological scope including ensemble data assimilation (Ensemble Kalman Filter, Ensemble Smoother with Multiple Data Assimilation), fuzzy logic and Type-2 fuzzy inference systems, physics-informed neural networks, polynomial chaos expansion, and Dempster–Shafer evidence theory. A generalized structure of IDSS with a dedicated uncertainty quantification module (IDSS_UQ) is presented, formally extending the classical DSS quintuple ⟨𝐷, 𝑀, 𝐾, 𝑈, 𝑅⟩ with the UQ module ℳ = ⟨𝛹, 𝛱, 𝛴𝑠, 𝛤⟩ and systematizing the integration of uncertainty characterization, propagation, sensitivity analysis, and calibration into the functional subsystems of IDSS. Unlike existing approaches that consider individual UQ methods in isolation, the proposed generalization unifies ten UQ method classes within a single structural specification aligned with the ISA-95 automation hierarchy. The placement of UQ modules within the ISA-95 (IEC 62264) automation hierarchy is specified, spanning edge-level real-time inference at Levels 1–2 through strategic analytics at Levels 3–4. Digital twin architectures with embedded UQ layers are analyzed, including deployments at BP, Equinor and Saudi Aramco demonstrating measurable production gains. The proposed methods are validated against five standard reservoir benchmarks (PUNQ-S3, Brugge, Egg Model, Norne, OLYMPUS) and mapped onto IDSS functional components. Results show that the combination of physics-informed surrogates with ensemble data assimilation achieves computational speedups of up to 6,000× over conventional simulation, while conformal prediction provides the only distribution-free finite-sample coverage guarantee applicable to petroleum time-series forecasting.
Zitieren
APA
Kytsmen, D. (2026). ARCHITECTURAL FRAMEWORK AND ADVANCED METHODS FOR UNCERTAINTY QUANTIFICATION IN INTELLIGENT DECISION SUPPORT SYSTEMS OF COMPUTER-INTEGRATED SYSTEMS FOR OIL AND GAS PRODUCTION. Measuring Equipment and Metrology, 87(2), 43-52. https://doi.org/10.23939/istcmtm2026.02.043
Vancouver
Kytsmen D. ARCHITECTURAL FRAMEWORK AND ADVANCED METHODS FOR UNCERTAINTY QUANTIFICATION IN INTELLIGENT DECISION SUPPORT SYSTEMS OF COMPUTER-INTEGRATED SYSTEMS FOR OIL AND GAS PRODUCTION. Measuring Equipment and Metrology. 2026;87(2):43-52. doi:10.23939/istcmtm2026.02.043
BibTeX
@article{Kytsmen2026,
author = {Kytsmen, Dmytro},
title = {ARCHITECTURAL FRAMEWORK AND ADVANCED METHODS FOR UNCERTAINTY QUANTIFICATION IN INTELLIGENT DECISION SUPPORT SYSTEMS OF COMPUTER-INTEGRATED SYSTEMS FOR OIL AND GAS PRODUCTION},
journal = {Measuring Equipment and Metrology},
year = {2026},
volume = {87},
number = {2},
pages = {43--52},
doi = {10.23939/istcmtm2026.02.043}
}
EUVIMED