Our staff has led and documented more than 10,000 PHA/HAZOPs of entire units in the past 20+ years. During those great many HAZOPs and What-Ifs and FMEA (we use various PHA methods as needed) we (the team) determine in the PHA meetings if we need a SIF and if so, what is the SIL necessary. We also decide when a SIF in not needed. So, in most cases, no further analysis is needed to determine the SIF & SIL necessary. (Of course, you still need to develop a detailed SRS, then design the system, and then do the SIL verification, and then install and validate — but those steps have nothing to do with SIF & SIL determination.)
So, contrary to what many think, a separate SIL review is NOT necessary and not required by standards. In fact, it is wasted effort and not desirable. Further, it definitely does not cause a failure of the PHA/HAZOP system. As I mentioned, we have done 10,000+ such studies and folks we have trained have done about 100,000 more such studies…. stating you do not need a SIF or that you need a SIF is very common and easy with the right HAZOP team and the right protocols. That is why the standards allow the same. But, if the PHA (e.g., HAZOP) team is confused on the risk, then LOPA is normally the next best review for that specific scenario; and from that review you may determine you need a SIF and then state the SIL needed. In general, companies under-focus on HAZOP reviews and over-focus on quantified reviews.
I’m one of the two originators of LOPA and was a primary author of the LOPA book from CCPS/AICHE (2001) and am primary author on the second related book, Guidelines for Independent Protection Layers and Initiating Events (2012); and we do a lot of LOPA since we helped develop it beginning in 1995; and we do some QRA and HRA as well. With that said, we are only dealing with “orders of magnitude” assessments and LOPA and QRA (bowtie, etc.) get overused. See the linked paper (FREE) for a related discussion:
Note that you will need to cut/paste the link into a browser since there are spaces in the file name… if you have trouble, drop me a line at:
or go to
and click on the file name.