We were recently asked about placing Hydrogen Sulfide (H2S) gas detectors in an area of the process where no H2S gas is normally present. Does the possibility for process upset conditions to abnormally result in the presence of toxic gas give rise to a requirement for detection? There was no way to simply answer the question, and I thought others could benefit from more perspective on the importance of gas detection philosophy. In many of these situations, there already was a Safety Instrumented Function (SIF) whose job it is to prevent toxic gas emission in the event of a malfunction of the air blower of a Sulfur Recovery Unit (SRU). If the blower fails, gas containing H2S can backflow into the air system and be routed to atmosphere under pressure to an elevated emission point. In this situation tolerable risk was already satisfied using that SIF and other IPLs to detect the blower failure and prevent the emission by closing isolation valves. Therefore, functional safety dictates that no additional risk reduction is needed from a gas detection system to mitigate the backflow; but if the tolerable risk target was not satisfied, then H2S gas detection might be advisable.
In absence of that contingency, we wouldn’t recommend gas detection on a risk basis. There is no firm performance-based requirement driven by functional safety. None the less, we have certainly seen some End User provide H2S monitoring in this area of the blower air intake for that exact concern.
Here’s why it makes sense for some and not for others….
When looking beyond the concept of tolerable risk reduction, the application of toxic gas detection as an instrumented safeguard becomes a matter of end user philosophy. Is gas detection there to detect credible (small) leaks that might reasonably occur that are attributed to mechanical integrity failures… corrosion erosion etc. ? These type of leaks are not otherwise protected by instrumented safeguards. So this is an example of a philosophy and it guides us in one direction. We would place gas detectors by evaluating credible leak sources in process services that normally contain H2S above a concentration threshold. These are events that will credibly occur in the lifetime of the process — and backflow wouldn’t be considered here.
At the other end of our the philosophical spectrum is major accident hazard mitigation. This is for large releases that would have the potential for life threatening impacts beyond the unit boundary. Fewer gas detectors might be needed because the events of concern are limited to the low frequency / high severity impacts. these major accident hazard scenarios should already be driven to tolerable risk via prevention — without the benefit of consequence mitigation such as gas detection and alarms . However, the end user might want to always have something on the “right hand side” of the bow tie diagram, which is where hazard mitigation resides. Under that philosophy, it makes sense to have consequence mitigation for most/all major accident hazards, and this would drive location of gas detectors, alarms and site emergency planning. One example of a major hazard scenario (among many others) would be the H2S emission scenario in the event of blower failure. Under this philosophy, we would indeed place gas detection here based only on severity rising to major accident hazard levels, not on likelihood. Not all end users would consider scenario a major accident hazard, but some would. The discriminator would be localized versus site -wide hazard or offsite hazard potential, and it is related to the concentration of H2S in the process and the pressure that drives the backflow & emission.
Regardless of the end user philosophy we would recommend consideration be given to a mapping study to give confidence that a release of concern will be detected by the type, numbers, location and configuration of toxic gas detectors. Contact Kenexis for further guidance or assistance.