Prediction of uncertainty quantification on the human consequences at industrial domino effect accident scenario / Anis Farhanah Mohd Suhaimi Yeong

A new and improved model for the evaluation of uncertainties related to the assessment of human consequences of a fire and explosion domino accident scenario is to be developed at the end of this research. While the scope of this research is limited to studying the impact of explosion on human, structural consequence is also an important aspect to be considered. This is because in the case of domino accidents, structural consequences such as structural failure and flying debris can play a significant role in causing damages farther away from the leaking area. Propagation of blast waves, or overpressure, lead to debris trajectories that could pose danger, should it hit any human within the vicinity of the trajectories. Consequently, this contributes to human impact, making it necessary to consider structural consequence in this study. In the case of liquefied petroleum gas tank explosion, the risk impact assessment on the storage facilities were carried out and a cylindrical LPG tank in Feyzin Refinery, France is selected as a point of study in this research. To obtain an accurate result of injuries, fatality and survival, the total burn surface area (TBSA) and age factor need to be considered, as their involvement would be a major contribution in predicting accurate level of injuries and probability of fatality to human. The results were compared and validated using several consequences model, such as the TNT Equivalent (BREEZE software), Baker-Strehlow model and ARIA investigation report. Software such as the Response Surface Methodology (RSM); the Mapping Application for Response, Planning, and Local Operational Tasks (MARPLOT); and Computer-Aided Management of Emergency Operations (CAMEO) were also utilized for a simplified approach in conducting the risk assessment. The simplified approach included age, TBSA and fatality inputs, where these inputs were then put into the RSM modelling. Based on the model validation result, the simplified model was able to identify the severity of injury experienced by receptors in the case of Feyzin accident.