Improving PHAs With Sample-Safeguards Checklists


This article represents the opinions of the author only. It does not necessarily reflect any official position of IIAR and is presented here for the consideration of the reader. Questions or comments about the article can be sent to IIAR and may be forwarded to the author at IIAR’s discretion.

For many ammonia refrigeration systems, a what-if/checklist process hazard analysis (PHA), with brainstorming about unique situations enhanced by a comprehensive safeguards checklist, provides a robust method for identifying hazards, analyzing risks, and developing any needed recommendations.

More collaboration on sample documents would help the industry complete better PHAs. The technical and legal tools for collaboration are improving rapidly, ranging from artificial intelligence to better revision-tracking software, and also including widely used copyright and licensing approaches.

Whether improving PHA methods for ammonia refrigeration is needed, and if so, how, has been discussed in the industry recently. For over twenty years, the what-if/checklist PHA method has
typically been used to evaluate refrigeration systems that contain 10,000 pounds or more of ammonia refrigerant in the USA. Some facilities have used hazard and operability (HAZOP) studies
instead, and the 2021 IIAR conference included a technical paper on “The Case for HAZOP,” by Stephanie Smith, P.E., Risk Management Professionals, Inc.

A full HAZOP study is appropriate for processes with unique science or engineering, which may include parts of some ammonia-refrigeration systems. Often ammonia-refrigeration systems are
designed to code, handbook, and manufacturer recommendations, and a good checklist covers most of the ground.

It would be very difficult to quantitatively compare PHA methods by what matters most — how much they improved safety — because:

  • multiple checks help ensure safeguards are in place (if a PHA team didn’t catch something, a building inspector, contractor, maintenance manager… might);
  • a PHA team’s effort may compensate for the pros and cons of the method they use, and;
  • ammonia refrigeration PHAs have often used blended methods, including some what-if questions resembling the methods of a HAZOP study (pressure, temperature, level… is high or low, here or there) and others resembling failure modes and effects analysis (this or that mechanical, electrical… component fails in some way).


One approach to improving a what-if/ checklist PHA is to start with a sample document that has:

  • fewer but broader what-if questions, to help the PHA team catch unique situations;
  • more detailed sample safeguards (also called controls), which serve as a checklist, and;
  • sample lists of causes and consequences, which may be modified as needed.

For comparison, based on a small sample:

  • some HAZOP and what-if/checklist PHA reports, for ammonia-refrigeration systems, have 300 to 400 scenarios, whose safeguards include 10,000 to 20,000 letters, approximately;
  • recent safeguard-checklist focused what-if/checklist PHA reports have 130 to 145 scenarios, whose safeguards include 50,000 to 65,000 letters, approximately.

With a longer sample-safeguards list, a big part of the PHA team’s effort becomes comparing sample safeguards to the planned or as-built system and deciding if any need to be implemented.

The goal is a PHA that’s both comprehensive and easier to complete.


The desire to avoid “canned” PHAs, that don’t get adequately customized, has led to under-appreciating the benefits of having sample cause, consequence, and safeguard wording, that
only get customized as needed. These benefits include:

  • a more comprehensive starting point (avoid missing well-known hazards and safeguards);
  • saving time and avoiding exhausting/ frustrating the PHA team (burnout);
  • easier to see what is unique about a refrigeration system or facility, which may create unusual risks, by using revision tracking to compare the customized to the sample PHA reports, and;
  • easier to aggregate the results across many facilities, to gain an understanding of which hazards and safeguards are widespread because the causes, consequences, and safeguards are semi-structured data

Semi-structured data is facilitated by giving sample causes, consequences, and safeguards both names and descriptions. The PHA team can then select these by name, modify the description as needed, but only modify the name if necessary. If an owner of many refrigeration systems then compares PHA reports between their facilities, they’ll see mostly the same names for causes,
consequences, and safeguards, with differences in their descriptions.


Some industries, via companies that are members of the Center for Chemical Process Safety, have started to compile anonymized PHA data from multiple facility owners to improve their PHA efforts. The cost and effectiveness of these artificial-intelligence approaches may depend on the quality of PHA reports and other documents used as inputs. This type of collaboration may allow
better comparisons of PHA methods, than the limited attempt made above, in this article.


IIAR standards provide a starting point for sample safeguards that PHA teams may review.

IIAR 9-2020, Minimum System Safety Requirements for Existing Closed-Circuit Ammonia Refrigeration Systems, calls for completing minimum systemsafety evaluations by about March 2025
and every five years thereafter. If a PHA is done in the meantime, shouldn’t it at least check if these minimum system safety requirements are met?

IIAR 6-2019, Inspection, Testing, and Maintenance [ITM] of Closed-Circuit Ammonia Refrigeration Systems, which is a required “normative” reference of IIAR 9-2020, describes ITM safeguards that a PHA team could compare to a facility’s ITM practices.

The design and building codes and standards that may apply to a refrigeration system and its supports, based on its location and installation start date (or building-permit application date),
likewise can serve as checklists. In recent decades, these usually include one or more IIAR-2 editions, the ASME and other codes & standards they reference, and any applicable building codes.

The contents of IIAR handbooks and guidelines, such as The Refrigeration Piping Handbook, could make a checklist too long, but a checklist can reference them, and the PHA team may
turn to them as needed, such as for information on avoiding shocks or upper limits for flow velocities in piping.

PHA teams should look further than the minimum requirements in standards. Safeguards identified by prior PHA teams over the years help do this. Some PHA leaders have compiled these, in documents or even just in their memory, to provide suggestions during PHA team meetings.

Collaborating on lists of sample safeguards, or entire sample PHA reports could aggregate this experience, from many PHA teams. Most of a PHA team’s effort, and hopefully the value they create, results from their evaluation of a specific facility and refrigeration system, producing customized recommendations. Collaborating on sample documents would help focus attention on this site-specific effort and would still allow competition between who best leads PHA teams to provide the needed customization.


If a goal is making sample PHA documents widely available and with changes well tracked, some free tools that could help include the following.

  • Git and GitHub allow tracking and sharing large sets of text documents. They also allow branching, for example, if two groups of people can’t agree on a change to a sample document, the document could be “branched,” with each group making their own version, and with the ability to view the differences between the branches as time goes by and more changes are made. The flexibility and informality of developing sample documents in this way would help distinguish them from standards, to which regulators may try to enforce compliance.
  • Legal tools for licensing sample documents (and computer code, etc.) that get modified for commercial use, which to date have protected contributors from liability, include the Apache, Creative Commons, MIT, and other licenses.