Dynamic risk assessment of CNG stations using the Bow-tie approach and Bayesian network

Abstract

Today, CNG refueling stations are expanding worldwide to provide vehicles with CNG as fuel. However, constructing CNG stations in urban areas has led to serious threats to public and property safety. Therefore, risk assessment in these locations is essential. In most cases, risk assessment is the most complex part of risk management. While static methods, such as the Bow-tie approach, pose major limitations to most conventional risk assessment methods, dynamic risk assessment offers a better picture of risks and their consequences. This study aimed to develop a dynamic and comprehensive quantitative risk analysis (DSQRA) approach to evaluate accident scenarios and model risks in CNG stations. In this approach, FMEA was used for hazard analysis, while a Bow-tie diagram and Bayesian network were employed to model the worst-case accident scenario and assess risks. According to the results, dispenser gas leakage was identified as the worst-case accident scenario, with hose cracking and corrosion being the most critical factor. Therefore, in the risk management of CNG stations, priority should be given to the most probable basic events and main contributing factors identified in this study to reduce the likelihood of accident scenarios and thus mitigate risks.