Cutting Corners in Procedural HAZOP

For many of us who facilitate Hazop studies, procedural Hazops come our way only rarely. Despite a decade or so of facilitating continuous process Hazops, I could count on the fingers of one hand the procedural Hazops that I’ve undertaken. It will happen, then, that we all get a bit rusty. Indeed, the reason procedural Hazops don’t come about more often may be that they appear so laborious as to not be worth it, compared to what could be discovered from a continuous Hazop of the same equipment. 

Just looking at the numbers, it’s easy to be put off. A typical set of continuous Hazop guidewords might contain some 20-30 items. A continuous Hazop might then cover perhaps 50 – 75 rows of node/guideword combinations in a day. 

For a procedural Hazop, the list is much shorter – maybe eight to ten guidewords – but these must be applied to each step in a procedure, which might run to dozens of steps. This can result in hundreds of rows of step/guideword combinations. Should we then expect procedural Hazops to take weeks? In practice, they needn’t be an overly time-consuming exercise. The trick is knowing what to cull. 

Many step/guideword combinations that arise in a procedural Hazop fail to add any insight. Slogging tediously through these will only lead the team to fatigue, boredom, and ‘silly season’. Instead, everyone in the room must be prepared to quickly skip over some rows, in search of the real hazards. For best efficiency, these may even need to be flagged for exclusion in advance of the Hazop workshop. 

Here are some examples of items that are not worth dwelling on: 

• Justifying the error. Procedural Hazop takes, as a given, that an error has been made in that procedural step. It’s quicker to review what would happen if that error occurred than it is to debate whether that error would occur (and then still have to review the outcome anyway). Accept that the error is possible and move on. 

• Nonsensical combinations. If the step says, “open the valve”, and the guideword is “reverse action”, then we should consider closing the valve instead. But by definition, it was already closed. This is a nonsense combination. Move on. 

• Repeat items. If we have already looked at Step 1, “out of sequence: after Step 2”, then it is unlikely anything new will be revealed by considering Step 2, “out of sequence: before Step 1”. These are exactly equivalent. The second time around, with the team now focusing on Step 2, it may be worth another quick look for any new thoughts – but otherwise, move on. 

• Hazoping check-steps. In continuous Hazop, we do not take a failed safeguard as a cause. In procedural Hazop, an entire step can be a safeguard, and in practice, there is little value in applying the guidewords to these. Don’t get bogged down with a step like “check that the controller is in Auto” (what, for example, would it even mean for a step like this to be done “too soon”?). Instead, let the earlier step of “put the controller in Auto” reveal the hazards of the associated errors, and move on. 

• Time-related guidewords, on steps with no time component. It’s just not possible for a step like “start the pump” to be done “too quickly” if, for example, it’s a fixed-speed direct-online motor. Move on. 

• Delay of operation. “Fill the tank”, plus “too late”, means you’re just sitting there waiting until you finally do remember to fill the tank. Yes, time is money, but a static situation (provided it is actually static) is not usually a hazardous situation. Move on. 

• Not-good-practice consequences. Opening or closing valves slowly avoids fluid hammer and shocks to the process and may provide an opportunity to immediately correct an error. As a result, the “too quickly” guideword usually at least makes sense. However, this is a generic point of good practice that more or less applies to all valves, all the time. If there isn’t a specific reason to worry about a step, move on. 

• Searching too hard for substitute actions. The target of the “substitute action” guideword is another action that could reasonably be confused with the intended one. If the procedure says to start the standby pump, for example, then we start the duty pump by mistake. We do not need to consider what would happen if, instead of starting the standby pump, an operator isolated the filter. Just. Move. On. 

This approach can seem anathema to facilitators more accustomed to continuous Hazop. Flagging guidewords in advance to skip over, and rushing the team through their brainstorming, would be terrible facilitation practice in a continuous Hazop. But for those of us who use both types, it’s important to recognise them as the two different methods they are and run them accordingly. 

Ultimately, the intent is to get the team to identify whatever hazard scenarios they can find. That requires a balance of pace and rigour, to keep everyone engaged. Because in the end, a room full of coma patients won’t identify any hazards at all.