A Novel Approach Based on Stochastic Hybrid Fault Tree to Compare Alternative Flare Gas Recovery Systems
2021
Flaring has always been an inseparable part of oil production and exploration. Previously, waste gas collected from different parts of facilities was released for safety or operational reasons and combusted on top of a flare stack since there was not the possibility to treat or use this type of gas. Concerns about global warming led to several initiatives for reducing flaring or even eliminating combustion. Treating flare gas was made possible by the introduction of flare gas recovery systems that have become increasingly obligatory. Most solutions add a flare gas recovery system to an existing flare system. In a typical scenario, after analyzing the existing facility and collecting the necessary data, alternative designs are proposed and criteria are determined to make a choice between the proposed alternatives. In this paper two designs of a gas control system are proposed, and reliability was chosen as the deciding factor. Using repairable dynamic fault trees, the failure models of the two designs have been implemented. Afterwards, a novel hybrid technique, the Stochastic Hybrid Fault Tree Automaton, is used to model the working conditions in which the system operates, with the aim to achieve a more realistic assessment and evaluate the disaster likelihood associated to these failures. It is shown that the latter enables a richer analysis where the effects of failure can be better assessed. This is important for correct choice between design alternatives because, as shown in the case study, the results of the two analyses can lead to contrasting conclusions of the solution to adopt. Further investigations have been carried out focusing on the safety sub-systems and on the basic events in each design. The Importance Measure analysis revealed that some of the components were responsible for most of the critical failures, thus locating some areas of possible design improvement.
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
49
References
1
Citations
NaN
KQI