A hazard and operability study (HAZOP) checks designs for safety risks. Unfortunately, companies too often broaden its role. As a fairly recent letter to the editor  pointed out: “An important rule for HAZOPs is that it is not a design activity, but rather a design check (for hazards and operability).” The letter writer was emphasizing that a HAZOP isn’t intended to solve the issue identified. Instead, later work, outside the HAZOP, should straighten out the problem. This makes sense. Trying to solve the problem in the HAZOP keeps the meeting going, tying up all participants for more time. Additionally, the participants in the HAZOP may not be the right people to resolve the issue. Of course, any changes proposed should be verified as well.
While the sentiment expressed by the letter writer is valid, my thinking is even broader. I strongly believe a HAZOP mostly is a safety design check, not a design development review. Making and finalizing major design choices should occur long before a HAZOP. Let’s look at some examples to see the difference.
In the first example, a client was well into the design process to build a new biosynthetic plant. Its output was to be a bio-based material that could serve as a drop-in replacement for an existing industrial product. Multiple parts of the process took place under vacuum. The vacuum systems selected used liquid-ring pumps. The HAZOP review identified the issue that maintenance on the pumps created personnel exposure to the seal fluid. One recommended action was to consider switching to steam-jet ejectors for the vacuum systems. This is the equivalent of having a HAZOP-recommended action of replacing distillation with freeze crystallization for purifying a product because the freeze crystallization runs at a lower (safer) temperature for personnel exposure.
The issue here is not whether you think liquid-ring pumps have better or worse safety than ejectors. Experience shows that both systems have well-understood reliability and safety characteristics. Rather, the issue is that revisiting a major technology choice is not the purpose of a HAZOP. Its aim is to deal with the safety and operability within the system you have.
HAZOPs occur relatively late in the design process. The place to make major technology choices is during preliminary engineering — or process hazard analysis (PHA), at the latest. Making fundamental changes in work scope late in the work flow has major consequences to cost and schedule.
In the second example, a set of HAZOPs was scheduled to review an entire new biofuels plant. By the time it was finished, this set of HAZOP meetings stretched over 18 weeks. About six weeks in, a new controls-and-instrumentation engineer showed up from the client company. In the very first session afterward, the engineer looked at the drawings and said “These interlocks don’t meet our company’s standards. You’ll have to redo them all.” Silence descended.
This started a month-long debate on what the standards really said and what actually was needed; it mostly involved arguments between various groups within the client company. I personally didn’t have an agenda on which approach to take; both looked acceptable to me. I just wanted a decision, any decision, but a decision — and to know which approach to use and whether everything needs to be redone. Again, this is a type of question that should be decided before a HAZOP.
If you ever hear “We’ll review that in the HAZOP” or “We’ll decide that in the HAZOP,” stop immediately. Work already is headed off the rails. The purpose of the HAZOP is to identify things to check, not to do engineering. That’s the point of the letter that started this discussion. And it’s certainly not to make conceptual changes, which is the focus of this column.
How you avoid late engineering is to get the right skills involved early. Well before any HAZOP there should be a PHA, a piping-and-instrumentation-diagram review, a construction review — or whatever else you think is needed. Too often, plants will not assign senior operating personnel to these early meetings. As a result, many operating issues get raised at the last minute. This is a project management error. Get the plant people involved earlier rather than later.