Well, that depends on what you call QRA. Some companies (the folks they have taught) define QRA as only the high level risk assessment that may be performed at the early design phase of a project to make large siting and fire protection (and other mitigation) decisions. LOPA does not replace that specific QRA. But that definition of the term QRA is far too narrow; we prefer to refer to those specific QRA as “Plot Plan QRAs”. But any name other than QRA that is descriptive of the goal will do.

CCPS defines QRA to be analysis that uses quantitative risk assessment methods such as FTA, ETA, HRA, and consequence modeling. That is the authoritative source (in the opinion of tens of thousands of chemical engineers) of the term QRA. Yes, some folks now loosely (in our opinion) use the term QRA to only refer to some worst case consequence and frequency modeling for use in fire risk assessments and mitigation design, facility siting, and a few other questions). the “Plot Plan QRA.” In reality, that single focus “QRA” is only using QRA methods (FTA, ETA, consequence modeling) for a specific business goal (e.g., siting, fire system load, etc.).

For those interested, please read the early LOPA papers (the two definitive papers from the mid-1990s by me and Art Dowell are free from our website, www.piii.com/resources/ ) Or, read the introduction chapter of LOPA (2001, CCPS/AIChE) for a nearly identical explanation (since Art and I wrote 70% or 80% of that textbook). Otherwise, it is easy to get confused on terminology and guessing about context…. those sources explain what this statement means: “LOPA replaces 90% of the QRAs.”

In summary, LOPA is only used for (only designed for) risk judgment of one scenario at a time. The scenario is chosen from a HAZOP, usually. LOPA helps a company make a judgment when the HAZOP team has difficulty doing so. But, a competent HAZOP (PHA) team can make a good risk judgment on 95 to 98% of the scenarios they uncover, without need of LOPA. THEN, for the few scenarios that need LOPA … are you listening?…. those scenarios formerly (pre-2001) went to FTA, ETA, modeling (e.g., the QRA methods in the QRA
textbook from CCPS and similar textbooks) to help make a risk judgment on a single scenario (the same scenario that can now use LOPA instead.). To date, LOPA has reduced the load by 99% of the single-scenarios going to QRA methods. But, since it is difficult to explain that concepts and get folks to understand, we hedge the factor and say that LOPA replaces 90% (though really it is 99%) of the scenarios that used to go to QRA methods instead (used to go to FTA, consequence modeling, etc.).

To learn more about the full scope of the term QRA, then first buy the book of Guidelines for Chemical Process Quantitative Risk Analysis, 2nd Edition, 1999. Then also study the published Probabilistic Risk Assessment (PRA) for Nuke Power plants… you will find out the full definition of QRA.

By the way, we sometimes do QRA on single-scenarios, when LOPA does not answer the question (there are many scenarios for which LOPA is not appropriate or not the best method). But we also do the High Level Risk Assessment (like the special case QRA Ahmed refers to) for the early design phase of a project.

Image Copyright © The Dallas Morning News