Advanced data-driven fault detection in gas-to-liquid plants

Nour Basha; Radhia Fezai; Byanne Malluhi; Khaled Dhibi; Gasim Ibrahim; Hanif A. Choudhury; Mohamed S. Challiwala; Hazem Nounou; Nimir Elbashir; Mohamed Nounou

doi:10.1016/j.compchemeng.2025.109098

Advanced data-driven fault detection in gas-to-liquid plants

Nour Basha, Radhia Fezai, Byanne Malluhi, Khaled Dhibi, Gasim Ibrahim, Hanif A. Choudhury, Mohamed S. Challiwala, Hazem Nounou, Nimir Elbashir, Mohamed Nounou^*

^*Corresponding author for this work

HBKU College of Science and Engineering

Research output: Contribution to journal › Article › peer-review

Abstract

Fault detection is a critical part of process monitoring, where the objective is to flag unexpected operating behavior quickly and accurately. In this paper, a novel extension of the Generalized Likelihood Ratio charts is proposed, denoted as the Maximum Multivariate GLR charts. Linear and nonlinear data-driven models, namely principal component analysis and its kernel extension and neural networks, are combined with different statistical charts towards the detection of multiple fault types in three distinct case studies: synthetic, Tennessee Eastman process, and Gas-to-Liquid process. The results show that the MMGLR charts have a better detection accuracy than conventional charts, and that neural networks are more robust modeling techniques than PCA and KPCA for the sake of fault detection.

Original language	English
Article number	109098
Journal	Computers and Chemical Engineering
Volume	198
DOIs	https://doi.org/10.1016/j.compchemeng.2025.109098
Publication status	Published - Jul 2025

Keywords

Fault detection
FischEr–Tropsch Synthesis
Gas-to-liquid process
Generalized Likelihood Ratio
Kernel Principal Component Analysis
neural network

Access to Document

10.1016/j.compchemeng.2025.109098

Cite this

@article{e5be561c8fe6481683fe849f509083f0,

title = "Advanced data-driven fault detection in gas-to-liquid plants",

abstract = "Fault detection is a critical part of process monitoring, where the objective is to flag unexpected operating behavior quickly and accurately. In this paper, a novel extension of the Generalized Likelihood Ratio charts is proposed, denoted as the Maximum Multivariate GLR charts. Linear and nonlinear data-driven models, namely principal component analysis and its kernel extension and neural networks, are combined with different statistical charts towards the detection of multiple fault types in three distinct case studies: synthetic, Tennessee Eastman process, and Gas-to-Liquid process. The results show that the MMGLR charts have a better detection accuracy than conventional charts, and that neural networks are more robust modeling techniques than PCA and KPCA for the sake of fault detection.",

keywords = "Fault detection, FischEr–Tropsch Synthesis, Gas-to-liquid process, Generalized Likelihood Ratio, Kernel Principal Component Analysis, neural network",

author = "Nour Basha and Radhia Fezai and Byanne Malluhi and Khaled Dhibi and Gasim Ibrahim and Choudhury, \{Hanif A.\} and Challiwala, \{Mohamed S.\} and Hazem Nounou and Nimir Elbashir and Mohamed Nounou",

note = "Publisher Copyright: {\textcopyright} 2025 Elsevier Ltd",

year = "2025",

month = jul,

doi = "10.1016/j.compchemeng.2025.109098",

language = "English",

volume = "198",

journal = "Computers and Chemical Engineering",

issn = "0098-1354",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - Advanced data-driven fault detection in gas-to-liquid plants

AU - Basha, Nour

AU - Fezai, Radhia

AU - Malluhi, Byanne

AU - Dhibi, Khaled

AU - Ibrahim, Gasim

AU - Choudhury, Hanif A.

AU - Challiwala, Mohamed S.

AU - Nounou, Hazem

AU - Elbashir, Nimir

AU - Nounou, Mohamed

PY - 2025/7

Y1 - 2025/7

N2 - Fault detection is a critical part of process monitoring, where the objective is to flag unexpected operating behavior quickly and accurately. In this paper, a novel extension of the Generalized Likelihood Ratio charts is proposed, denoted as the Maximum Multivariate GLR charts. Linear and nonlinear data-driven models, namely principal component analysis and its kernel extension and neural networks, are combined with different statistical charts towards the detection of multiple fault types in three distinct case studies: synthetic, Tennessee Eastman process, and Gas-to-Liquid process. The results show that the MMGLR charts have a better detection accuracy than conventional charts, and that neural networks are more robust modeling techniques than PCA and KPCA for the sake of fault detection.

AB - Fault detection is a critical part of process monitoring, where the objective is to flag unexpected operating behavior quickly and accurately. In this paper, a novel extension of the Generalized Likelihood Ratio charts is proposed, denoted as the Maximum Multivariate GLR charts. Linear and nonlinear data-driven models, namely principal component analysis and its kernel extension and neural networks, are combined with different statistical charts towards the detection of multiple fault types in three distinct case studies: synthetic, Tennessee Eastman process, and Gas-to-Liquid process. The results show that the MMGLR charts have a better detection accuracy than conventional charts, and that neural networks are more robust modeling techniques than PCA and KPCA for the sake of fault detection.

KW - Fault detection

KW - FischEr–Tropsch Synthesis

KW - Gas-to-liquid process

KW - Generalized Likelihood Ratio

KW - Kernel Principal Component Analysis

KW - neural network

UR - http://www.scopus.com/inward/record.url?scp=105000100362&partnerID=8YFLogxK

U2 - 10.1016/j.compchemeng.2025.109098

DO - 10.1016/j.compchemeng.2025.109098

M3 - Article

AN - SCOPUS:105000100362

SN - 0098-1354

VL - 198

JO - Computers and Chemical Engineering

JF - Computers and Chemical Engineering

M1 - 109098

ER -

Advanced data-driven fault detection in gas-to-liquid plants

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this