Using Fault Tree-Derived Bayesian Networks and Sensitivity Analysis for Probabilistic Risk Assessment of Quarry Accidents in Libya

Abstract

A Fault Tree Analysis-derived Bayesian Network forms the basis in this research for a probabilistic risk assessment framework that can be applied to life-threatening quarry events. The proposed model quantifies incidents such as fire and electric hazards, blasting hazards, haulage and machinery accidents, and slope failures. Conditional probability tables (CPTs) are built using data sourced from the expert input and consultation. The data are then input into GeNIe software to analyze sensitivity and inference levels. Analysis of these data shows that the main events contributing to quarry fatalities are haulage and traffic accidents (0.96), followed by machinery accidents (0.94) and blasting hazards (0.94). The proposed model contributes to the decision-making process in quarry environments by prioritizing safety, especially in data-scarce countries like Libya