Amos Necci
step 2) 5.2.3. Characterization of flood impact vector (step 3) 5.2.4. Model validation (step 4) 5.2.5. Dataset of failure conditions (step 5) 6 5.2.6. Simplified correlations for vessel damage (step 6) 5.3. Results and discussion 5.3.1. Analysis of vessel failure conditions 5.3.2. Sensitivity and uncertainty analysis 5.3.3. Assessment of vessel damage probability 5.4. Application to a case-study 5.5. Conclusions 6. Probability assessment of multilevel domino scenarios 6.1. Introduction 6.2.
more » ... tilevel domino assessment using Markov analysis 6.2.1. Methodology Overview 6.2.2. Preliminary considerations regarding domino scenarios 6.2.3. Secondary accident typology selection 6.2.4. Identification of targets for the escalation process and domino system definition 6.2.5. The accidental scenario set 6.2.6. System and "states" description: the construction of the Directed Acyclic Graph (DAG) 6.2.7. Probability of accident escalation: transition between states Equipment vulnerability due to several secondary scenarios Transition probability calculation 6.2.8. Application of the Bayes theorem: Conditional transition probabilities calculation and domino frequency assessment 6.3. Results 6.3.1. Application to simplified case study 6.3.2. Comparision with previous models 6.4. Conclusions Final conclusions Appendix Acknowledgements cascade or might be busy on finding a shelter. Finally, mitigation measures (e.g. containment dikes or foam systems) may not work properly due to the interference of the event cascade. However, due to the fact that accidents scenarios due to cascading events are characterized by a very high complexity, in combination with the extremely low probabilities of such accidents, those events are often left out from the safety assessment of chemical activities. For this reason, one of the main topics of academic and industrial research in the field of industrial safety is the study of cascading events, aimed at the development of specific methodologies for risk assessment and safety management. The activity of my PhD regarded cascading events, on the research of domino and NaTech events. In this thesis, a state of the art of available approaches to the modelling, assessment, prevention and management of domino and NaTech events is described. New methodologies, developed during my research activity, aimed at the quantitative assessment of domino and Natech accidents are also presented. Chapter 2 reports the state of the art on the research regarding domino accidents. In this chapter three main topics are analyzed: the analysis of past accidents, the development of fragility models for process equipment and the development of risk analysis and safety management methods. Chapter 3 reports the state of the art on the study of NaTech events. In this chapter a review of past accidents is reported, observed both in the US and in the EU. Then, the efforts toward a common framework for the quantitative assessment of NaTech accidents are described. Chapter 4 reports the efforts toward a novel methodology for the quantitative assessment of accidents due to lightning strikes. In this chapter contains: a detailed analysis of past accidents caused by lightning strikes, a statistical model for the assessment of lightning impact frequency on process units, the analysis and identification of possible accidents triggered by lightning strikes, a model for the assessment of process vessel vulnerability due to lightning strikes, the description of the protective barriers that protects storage tanks, event trees that helps the assessment of the probability of the final scenarios, a methodology for consequence assessment of accidents triggered by lightning strikes, a case study showing the result obtained by the use of the developed models to a real industrial installation. Chapter 5 contains the development of fragility models for the assessment of the damage probability of storage and process vessels during intense floods events. In this chapter, a mechanical model for the damage to horizontal storage tank is presented. The extensive use of the mechanical model lead to the identification of a dataset of failure conditions, allowing the statistical interpretation for the assessment of a damage probability for a given tank involved in flood events. Chapter 6 contains a novel methodology to describe and assess the possible scenarios that arise as a consequence of domino events. In this chapter the Markovian analysis is used in order to identify all the possible domino scenarios and their probability to occur. Chapter 7 contains the final conclusions of the work. This thesis presents the most up to date discussion, and uses the most advanced models, in order to provide the best methodologies to deal with cascading events, allowing their inclusion in safety management systems of the chemical and process industry. This study offers to be a milestone for the state-of-the-art for further research on the topic.