System hazard identification prediction prevention (SHIPP) methodology predictive accident modeling approach

A process accident occurs as a result of a sequence of events initiated by deviation in the process parameters and/or failure or malfunctioning of one or more components. Many process accidents are controlled and mitigated before they escalate to major events. Unfortunately some do go on to produce catastrophic consequence s. As the size and complexity of processing facilities increase, the potential risk posed by accidents is increasing. Operational safety could be improved by giving emphasis to the prevention of incidents, rather than relying on control and mitigative measures. This method is referred to as an "inherently safer approach ". To prevent major, though infrequent, event occurrence, it is important to consider accident precursors (symptoms of hazards) such as operational deviations, mishaps, and near misses, in order to prevent abnormal events at source rather than controlling or mitigating them. -- The objective of this research is to present a novel methodology known as System Hazards Identification, Prediction and Prevention (SHIPP) for process accident modeling and prevention. In this methodology, a new process accident model with predictive capabilities is developed. The SHIPP is a systematic methodology to identify, evaluate, and model the accident process, thereby predicting and preventing future accidents in a process facility. In this methodology, process hazard accidents are modeled using safety barriers. The model relies on process history, accident precursor information, and accident causation modeling. The fault tree and event tree analysis techniques are used to enhance the accident model and to represent a holistic picture of the cause-consequence mechanism of the accident process. Quantitative analysis has two aspects: updating and prediction. The model is able to capture the process operational behaviour, and update the accident likelihood using the Bayesian updating mechanism. The predictive model forecasts the probability of a number of abnormal events occurring in the next time interval. Application of this methodology is demonstrated by a case study. The quantitative results demonstrate that the probabilities of abnormal events dramatically change over time as new information is observed, and the adequacy and accuracy of model prediction is better in short term prediction rather than long term prediction. -- Through the SHIPP methodology, qualitative and quantitative analyses provide insight to identify critical safe ty barriers and functions, and determine the likelihood of failure of these measures. Combining management oversight, human factor and engineering analyses, the SHIPP methodology provides a comprehensive, systematic approach to manage a process system risk