The design of relief systems, specifically relief valve sizing, remains a crucial factor in mitigating the consequences of overpressure accidents in chemical plants. General criteria from engineering standards (API and ASME) and established engineering practices of operating companies and engineering contractors have provided guidelines in sizing relief valves. Sizing relief valves using standard and independent scenarios such as external fire, blocked outlet, and cooling water failure is a straight-forward matter however in instances where double contingencies or double jeopardy would apply, it is increasingly more difficult to assess the scenarios.
The paper will provide the general pathway to determine quantitatively if a double jeopardy scenario should become credible by using a deductive failure analysis method called Fault Tree Analysis. Fault Tree Analysis can analyze the overpressure state of the system using Boolean logic to combine lower level events such as two separate failures in a double contingencies scenario. A case study for seven permutations of control valve failure and its bypass valve positions will be compared and discussed with reference to Fault Tree Analysis results. Three of the permutations are based on engineering practices from several of operating companies and an assessment is made as to which approach is more realistic for design.
Fault Tree Analysis has been selected for study since it is considered an appropriate mathematical tool for analysis. By taking into account operation and design experience, Fault Tree Analysis can assist design engineers to incorporate or dismiss certain double jeopardy scenarios.