“Come on, baby, let’s start anew, ‘cause breaking up is hard to do.”  — Neil Sedaka

Before a HazOp team ever assembles, the facilitator has some important tasks to complete: get agreement with management on the scope of the review, identify the boundaries of the study, and break the process up into nodes.  And breaking up is hard to do.

Guiding Approach

The original guidance for identifying nodes in a process HazOp was that a node was either a vessel or a line connecting vessels.

That was good as far as it went, but it wasn’t very nuanced.  What about branches in lines? Jackets on vessels? Pumps? Accessories, like drains, vents, and impulse lines to pressure sensors?

Then there was a whole new school of thought that emerged, that a HazOp Node was a system. By system, some practitioners would include several vessels, several lines, and extend their node across multiple P&IDs.

One of our clients makes this statement in their policy and procedure for HazOps: “Large or small is not better—the size of the node must be just right.

Okay! “Just right.”  What is that?

Too Big

The system approach considers entire systems as single nodes. The cooling water system includes the cooling towers, the cooling water pumps and distribution headers, the cooling water supply lines, the cooling water return lines, the cooling water return headers, the make-up water, the tower blowdown, the water treatment chemical supply lines, and the water treatment chemical tanks. These each have their own hazards, and to capture each one in a single node is a mighty challenge.

Likewise, a distillation system includes the feed, the distillation tower itself, the overhead line, the overhead condenser, the reflux line, the tower bottom line, the reboiler, the reboiler line, the overhead takeoff, the bottom take off, any surge tanks, and any mid-column take-offs. Even the deviation of “Level-high” is ambiguous in such a case and will require being addressed many times.

Too Small

The most excruciating HazOp will be when every section of the process, from one fitting to the next, is considered a node. Every branch in a line, every vent, every drain, every level loop. No experienced facilitator advocates this, and no new facilitator needs to feel any need to do this.

There is one comment to be made in defense of small nodes. One of the responsibilities of a facilitator when preparing for a HazOp is to set up worksheets for each HazOp node. There is no reason to tie up the valuable resource of HazOp team in performing an essentially clerical function that does not require their input. Yet during a HazOp, the team will occasionally disagree with how the process was divided, usually because of a misunderstanding of the process by the facilitator.  When the node is too big, it takes time to set up a new worksheet. When the node is too small, however, it takes no time at all to simply state “Node 10 considered with Node 09.”

Still, it is preferable to get it just right.

Just Right

Nodes typically consist of vessels or lines.  When two separate parts comprise a vessel, such as the shell side and the tube side of a heat exchanger, each part is a separate node, because differences in composition, pressure, or loss of containment change the nature of the hazards in each part.  The same holds for a jacket and the vessel it jackets.

An entire line is a node.  When a line includes a pump or a blower, the pump is included as part of the node.  Depending on the nature of the line, it also may be appropriate to include branches of the line as part of the node, particularly where the composition, pressure, temperature, or loss of containment pose the same hazards.

Whether lines or vessels, nodes will typically include accessories, like vents to atmosphere, drains to grade, bypasses around control valves, sample lines, and impulse lines to pressure sensors, unless there are compelling reasons to treat the accessory as a separate node. One compelling reason is when vents are not to atmosphere or drains are not to grade, but to headers, where back flow can become a hazard.

A Study

A few years ago, we had a client that was interested in question of node size. It was common for them to conduct their HazOp based on systems—the large node approach. We explained that we used smaller nodes—the vessel or line approach. We were all curious to see which was better. If the larger node approach was more effective, we were going to switch, not just for them, but for all clients.  It would make us more competitive.  If the smaller node approach was more effective, they were going to convert for all HazOps in the future. It would make them more competitive.

Together, we conducted HazOps on two processes.

Each process was reviewed twice. One facilitator broke the two processes up into large nodes. A second facilitator broke the two processes up into smaller nodes. Then the two facilitators, working together, led the four HazOps.

The smaller nodes gave two improvements. The first is that the reviews took about 20% less time. We expected that it would because of opportunities to take advantage of duplication.  But it also happened that the team spent less time circling back around and rehashing as it worked through the process. Perhaps more importantly, the reviews caught more hazards. There were no instances where the larger node approach caught hazards that the smaller node approach missed, but several instances where the smaller node approach caught hazards that the larger node approach missed.

It’s a Matter of Judgment

Not many organizations have the opportunity to conduct a study comparing approaches. Usually, the approach is based on the feelings of the facilitator and what the members of the team are used to.

Keep in mind that either approach requires judgment on the part of the facilitator of “what to leave in, what to leave out.” We all have deadlines and commitments. Because of the judgment required, the division of a process into nodes will typically be similar from one analyst to another when they are using the same approach, but differences should be expected.

While the two approaches differ, they are both systematic and both will discover hazards. In the words of anxious parent’s everywhere, “make good choices.”


  • Mike Schmidt

    With a career in the CPI that began in 1977 with Union Carbide, Mike was profoundly impacted by the 1984 tragedy in Bhopal and has been working on process safety ever since.