Monday, April 11, 2016
Exhibit Hall E (George R. Brown )
Chemical process facilities are moving toward the use of performance-based design techniques for fire & gas detector placement. These approaches are displacing methods for layout of gas detection based on uniform "grids" that date back more than 25 years. First generation performance-based design techniques utilized ‘gas volume detection’ principals that are valid in terms of detecting and mitigating a large explosion event, but can result can result in an excessively large number of detectors and a costly design. This paper presents a superior alternative to the first generation technique of volume detection that uses sophisticated modeling of gas release scenarios with computational fluid dynamics (CFD). The techniques described in this paper align with ISA TR 84.00.07 guidelines for scenario-by-scenario basis. Furthermore, the paper breaks the paradigm that CFD can only be used sparingly on complicated gas dispersion problems. With the proper definition of the process, optimal selection of boundry criteria, and automated scripting of multiple scenarios and data collection from those scenarios, the thousands of CFD runs required to develop results can be efficiently and cost effectively performed. This paper will demonstrate practical application of CFD modeling and probabilistic risk analysis in order to improve on the performance of gas detection systems while at the same time, reducing unnecessary conservatism in the number of detectors. The paper will present methodology for determining scenario-based gas detector coverage and include a worked application example for an offshore gas processing platform.