It seems to be that everyone is in agreement that there needs to be more consistency in how SIS are identified. It only makes sense that a standard process unit, such as a Hydrotreater in a refinery, should have standard functions that are always applied to the unit, and the SIL target for those functions should usually be the same. I’ve often heard it lamented that this is not the case, and it was highlighted at the Safety Lifecycle Management panel discussion at the Emerson Exchange. The solution that many propose to fix the problem is better and more elaborate “processes” for identification. Their solution is bigger more elaborate rule sets for their LOPA .
I could not disagree more! What they are trying to do is to have “process” replace “experience”, and in my experience, that fails every time it is tried. I have seen far too many LOPA’s done by the leading experts and luminaries fail miserably to identify SIF that are patently obvious to instrument and control engineers with only a modicum of experience.
When you think about it, consistency in SIS was not a problem before the release of the ISA 84 standard. The reason was that process licensors and engineering companies “knew” what SIF were required based on prior experience with the plant or the unit operations that it contains. Now that experience is being thrown out the window in favor of a “process” (i.e., HAZOP/LOPA), those SIF are being left out because the HAZOP/LOPA process didn’t “fall into” it.
The real solution lies in a concept that was explained during the Project Services Roadmap at the Exchange, and that Kenexis has been using almost since our founding. The solution is “Smart Process Solutions” as Emerson called them, or “Template Designs” as we call them at Kenexis. When Kenexis analyzes that SIL requirements of a process plant, we don’t need the HAZOP to hopefully stumble into the right SIF, because we already know what they are based on our long experience. Not only do we already know what they are, but we already have pre-completed template analyses that include the list of expected safety instrumented functions with a full SIL selection analysis including the hazard each SIF is intended to protect against, the initiating events that could cause the hazard, the non-SIS protection layers that are typically employed, and the resultant SIL target that we typically see. Using these templates as a starting point, and customizing for the specific process unit at hand is not only a better way at acheiving consistency, than “hoping for the HAZOP”, I would argue that ignoring this experience borders on negligence. In the next few weeks you should see some of these “Smart SIS Solutions” showing up on our web site.