(106l) Dynamic Operational Risk Assessment with Bayesian Networks
- Conference: AIChE Spring Meeting and Global Congress on Process Safety
- Year: 2012
- Proceeding: 2012 Spring Meeting & 8th Global Congress on Process Safety
- Group: Liaison Functions
- Time: Tuesday, April 3, 2012 - 5:00pm-6:30pm
Oil/gas and petrochemical plants are complicated and dynamic in nature. Ageing of equipment/components, season changes, stochastic processes, operator response times, inspections and testing time intervals, and timing of safety system operations are time dependent criteria that can influence the dynamic processes. But conventional quantitative risk assessment techniques have limited ability to address these kinds of time dependent effects. These methods cannot describe the variation of operational risks as time-dependent deviations or changes happening in the process. For example, the fault tree/event tree methodology provides a static logical relationship between a component output event and its failure or its effect on another component output event. But it is not able to reflect the time-dependent effects. Therefore, risk assessment methodology that integrates time factors for dynamic processes is very much important.
Dynamic Operational Risk Assessment (DORA) is a risk assessment technique proposed for complicated and dynamic processes which treats the testing/inspection interval and repair time as a critical parameter for evaluating the operational failure . The probabilistic modeling part of DORA consists of stochastic modeling and process dynamics modeling. DORA considers abnormal events to evaluate the scenarios and is not dependent on conventional quantitative risk assessment methodology. It is assumed that a component can have three states in its lifetime: normal operating state, abnormal event undetected, and abnormal event detected and repaired. The time a component stays in a state is defined as the sojourn time and for sojourn time distribution, it is assumed that the component failure rate will remain constant for its lifetime. The Monte Carlo simulation technique for system-state trajectory provides ideas for system state transitions and the effects of testing/inspected interval on the probability and frequency of the component’s abnormal events over a prolong time period have been studied.
A study by UK Health and Safety Executive  also reported ageing as a significant factor in process industry accidents. Corrosion, build up of dirt and debris, fatigue, etc., can affect the ageing of a component. Studies have identified that the component failure probabilities significantly increases with an increase in ageing of the components, if effective tests and maintenance is absent . Therefore, effects of ageing should be considered. In the DORA study, the component failure rate is considered constant and remains the same value for all component ages. But as it is described, it is important to consider the effects of ageing on failure rates. Therefore, the objective of this paper is to present how the ageing of equipment and components can be integrated to assess operational risks within the framework of DORA. A system state trajectory simulation is performed considering the component’s failure rate as a function of component’s increasing age. The simulation is done using matlab computer programming. The effects of the testing/inspection interval are also evaluated. The age dependent DORA study provides system degraded behavior due to the effects of ageing on abnormal events of a component. A case study on commonly used oil/gas separators on offshore plants illustrates this model.
1. Xiaole Yang and M. Sam Mannan; The development and application of dynamic operational risk assessment in oil/gas and chemical process industry, Reliability Engineering and System Safety 95 (2010) page- 806-815
2. UK Health and Safety Executive; Plant ageing study-phase 1 report, 2010
3. W. E. Vesely, Approaches for Age-Dependent Probabilistic Safety Assessments with Emphasis on Prioritization and Sensitivity Studies, NUREG/CR-5587 SAIC-92/1137 RG, R1, R9, August 1992
This paper has an Extended Abstract file available; you must purchase the conference proceedings to access it.
Do you already own this?
Log In for instructions on accessing this content.
|AIChE Graduate Student Members||Free|
|AIChE Undergraduate Student Members||Free|