Sensitivity Studies of Offshore Gas Detector Networks Based on Experimental Simulations of High Pressure Gas Releases

A quantitative assessment on the placement practices of gas detectors in the process industries

“…To address these issues, the common recommendation is to further reduce the spacing between detectors. In the case of Kelsey et al (2002Kelsey et al ( , 2005 this resulted only in a slight decrease in the mean detection times.…”

“…When additional layers of detectors are used at several heights, not only absurd amounts of detectors are required, but the results achieved still continue to underperform those of the RA. Kelsey et al (2002Kelsey et al ( , 2005 have previously assessed the performance of this approach on a simulated offshore facility, their findings outline to a great extent the reasons for the poor performance observed in Figs. 1e8.…”

“…Additionally, our results provide quantitative evidence of a well-known deficiency of this approach. The number of detectors necessary, even for the standard 5 m cloud, can be prohibitive (CCPS, 2009;ISA, 2010;Kelsey et al, 2002Kelsey et al, , 2005. Since this methodology was initially developed for offshore facilities, this is a special concern for onshore facilities, which usually have even larger volumes.…”

“…That is, identifying accidental releases as fast as possible, so that proper countermeasures can be initiated (IEC, 2007;NORSOK, 2008;ISA, 2010). This point of view is shared by recent performance analyses where gas detection systems effectiveness is commonly evaluated in terms of time to and probability of detection (Kelsey et al, 2002(Kelsey et al, , 2005Bratteteig et al, 2011). Gas detection systems have interfaces with the Emergency Shut Down (ESD), Blow Down (BD), Ignition Source Control (ISC), ventilation, Public Address (PA) and alarms system, and fire fighting systems (NORSOK, 2008).…”

Optimal gas detector placement under uncertainty considering Conditional-Value-at-Risk

“…Rigorous simulation of gas dispersion is needed to accurately calculate these outputs, and several key variables such as process conditions and geometry, leak locations and gas properties, and weather conditions all have significant impact on these dispersion simulations. The influence of these key variables is reviewed in the works by Bratteteig, Hansen, Gavelli, and Davis (2011), Kelsey, Hemingway, Walsh, and Connolly (2002), Kelsey, Ivings, Hemingway, Walsh, and Connolly (2005), and Marx and Cornwell (2009).…”