Integrated Advanced Risk Prioritization of Major Hazards in Oil and Gas Construction Projects - A Case-Based Study FMEA–AHP and Bow-Tie Analysis
Oil and gas construction projects present a compound risk environment in which conventional construction hazards interact with hydrocarbon-process interfaces, temporary works, lifting, excavation, electrical activities, hot work and simultaneous operations (SIMOPS). This paper develops and demonstrates an integrated risk-prioritization framework that retains the practicality of a conventional probability–severity risk matrix while adding failure-mode analysis, explicit multi-criteria weighting and barrier-based Bow-Tie analysis. A case-based analytical study was undertaken using hazard and control information contained in supplied Oil and Gas infrastructure Limited HIRA and process-safety documents, together with publicly disclosed company safety information. Twenty major hazard scenarios were structured into a comparative dataset. Traditional risk was calculated as probability multiplied by severity. Failure Mode and Effects Analysis (FMEA) was then applied using severity, occurrence, frequency and detectability, followed by an Analytic Hierarchy Process (AHP) weighting scheme of 0.40, 0.25, 0.20 and 0.15 respectively. The resulting normalized integrated index, termed the Fuzzy/FMEA-AHP-inspired Risk Priority Index (FRPI), was used to rank hazards and guide Bow-Tie barrier selection. The analytical results placed excavation collapse, suspended-load struck-by, SIMOPS/communication failure, work at height/man basket and buried-utility strike among the highest priorities. Illustrative control scenarios reduced the baseline P×S scores by approximately 33–67%, depending on the hazard. The study demonstrates that an integrated approach can discriminate between hazards that receive similar traditional risk scores and can translate risk ranking into auditable preventive and mitigative barriers. The numerical expert-rating dataset is explicitly treated as an analytical demonstration dataset where authenticated field ratings were not present; therefore, the paper does not claim that these illustrative values are measured site observations. The framework is intended for validation through an authenticated project HSE expert panel before operational adoption.