The Best Order for Hazards Analysis Recommendations

Thoughts on a Process Safety Beacon

Credit: Center for Chemical Process Safety

The May 2023 edition of the Process Safety Beacon describes an incident in which a truck driver unloaded sulfuric acid into a sodium hypochlorite tank. The error resulted in the release of a cloud of chlorine. Many people required medical treatment.

The recommendations in the Beacon were to do with,

• Pipe labeling,

• The correct marking of connection points,

• The need for correct procedures, and

• Discussing the loading and unloading of chemicals during a hazards analysis.

These recommendations are good, but they miss the potential to remove risk rather than merely reducing it.

Credit: Ian Sutton

Some years ago I managed a process hazards analysis that ran across a similar situation to that described in the Beacon. The facility had two small tanks: one for acid and the other for ammonia. (The chemicals were used for pollution control.) The loading lines were as shown in the sketch, and they had similar flange connections.

This part of the facility was new. The first time a truck showed up to fill one of the tanks the driver connected to the wrong tank connection — and it’s not hard to see why. The loading lines crossed one another as shown.

The hazards analysis team made recommendations such as we see for the Beacon incident. Our recommendations included switching the loading points, labeling the lines, and installing connections that were specific to the chemical being unloaded.

But the client was more perceptive than the team members

The client accepted our recommendations, and acted on them. But then he moved one of the tanks to the other side of the facility.

The lesson to be learned here is foundational.

The only way to eliminate risk is to remove the hazard

Risk has three components: a hazard, the consequences of that hazard should it occur, and the likelihood (not probability) of occurrence. These elements can be connected as shown in the following equation.

Risk (Hazard) = Consequence^^n * Predicted Frequency

When a hazards analysis team identifies a risk that is unacceptable the team is often asked to make a recommendation. The general tendency is to address the above terms in reverse order. First, they look for ways of reducing the likelihood of the event. In the case of the Beacon example, they suggest better labeling and procedures.

Having identified means of reducing likelihood, the team is then likely to suggest means of mitigating the consequence of a potential event. The Beacon does not have recommendations regarding consequence control, but examples could be the installation of a deluge system or ensuring that spills are sent to a dedicated drain system.

But the best way of reducing risk — and the only way of eliminating risk — is to remove the hazard. Which is what our client did. By moving one of the tanks to the other side of the facility he made it nearly impossible for the tank driver to inadvertently connect his vehicle’s hose to the wrong tank.

The lesson is as follows — if risk has to be reduced, most process safety teams will make recommendations in the following order,

1. Reduce the likelihood of an event,

2. Reduce the consequences of the event, then

3. Eliminate the hazard.

In fact, it will generally be more effective to tackle these terms the other way round. In particular, teams should look for ways of removing the hazard so as to reduce risk to zero.

---

The writing of this post suggests four more posts to do with risk and hazards analysis:

1. Thoughts to do with the vexed topic of “acceptable risk”.

2. Whether hazards analysis teams should be making recommendations.

3. Why the consequence term in the risk equation has an exponent whose value is greater than one.

4. The distinctions between frequency, predicted frequency, likelihood and probability.

---