Pipeline failure due to various threats may contribute to significant adverse impact on human safety, environment, and economy. In order to mitigate the severity of pipeline failure consequences, maintaining the integrity of the vast and aging pipeline structure is crucial. The main concern in offshore risk analysis is the unpredictable and uncertain pipeline conditions. In probability theory, Bayesian network is known as a powerful tool for knowledge presentation and condition inference under uncertainty. Probability analysis of pipeline damages is necessary to prevent unwanted incidents that may cause catastrophic accidents. In this paper, a Bayesian network model was developed to identify and analyze the probability of subsea pipeline condition subjected to corrosion, third party, operational, and material damages. Statistical data and experts' knowledge were integrated in addressing data limitation. Attaining the proposed network model, diagnostic analysis, mutual analysis, and sensitivity analysis were performed to validate and provide a substantial amount of confidence on the outcomes of the proposed model. The analyses have demonstrated that estimation of the developed model is reliable. The outcome obtained can be used to assist the decision maker to prepare preventive safety measures and allocate proper resources to significantly minimize the occurrence probability of the risk factors.
The authors would like to express appreciation for the support of thesponsor Universiti Malaysia Pahang Internal Grant [RDU1703169].
- Bayesian network
- probability analysis
- risk assessment