The United States Environmental Protection Agency in Washington DC. (Photo by: Loop Images/UIG via Getty Images)

By R. Lee Vail, P.E., PH.D.

On March 14, 2016, Environmental Protection Agency (“EPA”) proposed changes to the Risk Management Plan Program (“RMP”) Rule. On December 21, 2016, the EPA disclosed its changes via a Pre-Publication Copy.  A series of blogs are planned to address five major changes: root cause analysis for near misses, third-party audits, inherently safer technology, emergency response, and availability of information. This blog addresses the first revision: root cause analysis for near misses.

The actual language of the incident investigation requirement has not changed significantly as a facility always was required to “investigate each incident which resulted in, or could have resulted in a catastrophic release of a regulated compound.” 40 CFR 68.81(a). The revised rule basically says the same thing except that it describes the concept “could have resulted in a catastrophic release” by relating said comment to an undefined term: “near misses.”[1] Citing its own 1996 response to comments, according to the EPA, “the range of incidents that reasonably could have resulted in a catastrophic release is very broad and cannot be specifically defined.” The EPA declined to add a definition in this rule making. So what is a near miss?

EPA appears to rely on a vague Center for Chemical Process Safety (“CCPS”) definition that describes a near miss in terms of an event that could have resulted in catastrophic release “if circumstances had been slightly different.” With this as a starting place, the EPA provided several examples of near misses:

  • a runaway reaction that is brought under control by operators is a near miss that may need to be investigated to determine why the problem occurred, even if it does not directly involve a covered process because it may have led to a release from a nearby covered process or because it may indicate a safety management failure that applies to a covered process at the facility;
  • fires and explosions near or within a covered process;
  • any unanticipated release of a regulated substance; and
  • some process upsets.[2]

If not completely clear by the examples, the EPA added that “near misses should also include incidents at nearby process or equipment outside of a regulated process if the incident had the potential to cause a catastrophic release from nearby regulated equipment.” This could result in mandatory RMP incident investigations of utility systems such as power and steam generation.

So what isn’t a “near miss?” Unfortunately, all we have from the EPA is “the intent is not to include every minor incident or leak, but to focus on serious incidents that could reasonably have resulted in a catastrophic release, although EPA acknowledges this will require subjective judgment.” Regardless of examples, the salient point is that the event could have resulted in a catastrophic release. Is that not what the original rule said? As such, lifting of a relief valve that is routed to a properly designed and operated flare would not be expected to have reasonably resulted in a catastrophic release. Some processes are inherently unstable and result in frequent “runaway reactions.” In these cases, the process design probably contains multiple layers of protection. Using a layer of protection, by itself, should be insufficient to declare a near miss.

I also question the EPA’s sole reliance on subjective judgment. “Minor releases” could be objectively defined (e.g., less than a reportable quantity). Whereas a Title V deviation may occur if a control device operates at 99% efficiency instead of the required 99.5%, such can objectively be shown not to result in a possible catastrophic release.

One thing is clear; facilities need to consider, document and follow subjective and objective criteria to define incidents that could have resulted in a catastrophic release. Examples provided by the EPA should be appropriately addressed. Ultimately it may add clarity to minimize use of the undefined term “near miss” as it is ultimately introduced in terms of an incident that could have resulted in a catastrophic release (which is nothing new).  Also clear, the new rule did anything but add clarity.


[1] To further confuse the issue, the EPA related the undefined term to the concept using the Latin based phrase “i.e.,” which is translated as “namely” or “that is to say” and then states within the preamble, that “near miss is an example of an event that could have reasonably resulted in a catastrophic release.” Typically the Latin based term “e.g.,” is used to introduce examples. Is an event that could have reasonably resulted in a catastrophic release and example of a near miss or is that the de facto definition?

[2] The EPA further cites the CCPS concerning the investigation of process upsets to include: excursions of process parameters beyond pre-established critical control limits; activation of layers of protection such as relief valves, interlocks, rupture discs, blowdown systems, halon systems, vapor release alarms, and fixed vapor spray systems and activation of emergency shutdowns.