Abstract
The evaluation of subsea pipeline conditions and the calculation of the likelihood of failures are among the important factors for effective maintenance decision-making. Traditional qualitative methods to calculate the likelihood of failures are subjective, highly dependent on the experience and knowledge of the decision-makers, and suffer data limitations. Thus, the calculated likelihood of failures may not reflect the actual value, resulting in an improper maintenance program. In this work, an analysis of subsea pipeline conditions based on a Bayesian Network was proposed to handle knowledge uncertainties and assist in decision-making. This work aims to elucidate the conversion of experts' perceptions into a pseudo-quantitative likelihood for conditional probability tables (CPTs) elicitation for the proposed Bayesian network. The experts' opinion was transformed into a more crisp value to be integrated with the objective data for accurate determination of the failure likelihood. The formulation to predict the likelihood of pipeline failures that relies on experts' perceptions was developed using the artificial intelligent fuzzy analytical hierarchy process (FAHP) with the decomposition method. The proposed pseudo-quantitative formulation was established and was able to complement the existing risk-based model, which enabled the making of more informed pipeline maintenance decisions. The approach assisted the experts in eliciting the probabilities of nodes with emphasis on generating the conditional probabilities of the nodes with multiple parents.
Original language | English |
---|---|
Pages (from-to) | s187-s196 |
Number of pages | 10 |
Journal | Process Safety Progress |
Volume | 41 |
Issue number | s1 |
Early online date | 2 Mar 2022 |
DOIs | |
Publication status | Published - Apr 2022 |
Event | 2nd International Conference and Exhibition Loss Prevention Asia - Seri Iskandar, Perak, Malaysia Duration: 22 Nov 2021 → 23 Nov 2021 http://caps.utp.edu.my/index.php/loss-prevention-asia-2021/ |
Bibliographical note
Funding Information:This work is supported by the Ministry of Higher Education Malaysia under the Fundamental Research Grant Scheme (FRGS) project, FRGS/1/2019/TK02/UMP/02/27 (UMP reference: RDU1901206).
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.Keywords
- conditional probability table
- decomposition method
- fuzzy analytical hierarchy process
- knowledge uncertainties
- pseudo-quantitative likelihood