Some straight talk on process safety
Will increased connectivity change the definition of a process?
Plant Engineering recently sat down with Adam Kern, an engineer at Matrix Technologies, Indianapolis, about his work as a process safety engineer. Following are some of Kern’s insights into an important manufacturing function.
Question: When it comes to safety audits, what are some common issues encountered? What are, potentially, the most serious issues?
Adam Kern: The most common issue I see is related to time intervals and due dates for recurring actions. Some intervals are defined by the PSM (process safety management) standard, such as PHA (process hazard analysis) revalidations need to occur every five years. Those tend to be well understood and someone makes sure those get done on time. Other intervals are self-imposed, such as weekly checks done by operators. Those types of intervals are sometimes not taken seriously, and records show that lot of times. OSHA and the EPA will assume that if a task is important enough that a company says they are going to do it, then any evidence that they are not doing it may be worthy of a citation. For self-imposed intervals especially, I think it is important for a company to evaluate what is truly needed, and to commit to those while leaving unimportant tasks out. That allows you to focus resources on important things and not overload people with tasks that sounds good but don’t really add value.
The most serious findings that I’ve seen that are also pretty common are related to incident investigation, especially concerning near misses. Very often, I see documentation of why a near-miss was not an incident, rather than why it could have been. That’s because if you call it an incident, it requires documentation and people want to avoid that. That can cause things to be ignored that shouldn’t be, until they become serious enough to be a big incident, rather than preventing the incident in the first place. It is a difficult problem, because it requires the operations team, principally tasked with recognizing there an incident occurred, to react appropriately despite having of other duties in the moment.
Question: If someone ever finds themselves part of an incident investigation, how should they proceed?
Adam Kern: If you are being interviewed as part of the investigation, be as honest as you can. If you know that it is coming but the interviewer hasn’t gotten to you, write down as much as you remember after it happens, because it is easy to forget details the further removed from the incident that you are.
If you are leading an investigation or gathering information to support it, make sure that you are not accusatory of anyone as you go through. If people feel like they are being targeted for making a mistake, they will quickly clam up and not give you vital information for determining root causes. The majority of the time, the root cause of the incident is not simply human error, and more often can be traced to a systematic problem such as poorly written procedures, lack of enforcing procedure usage, or bad engineering design. Blaming a single individual is very rarely the correct path and generally fixes nothing.
Question: You mentioned one current regulatory-related issue is that of determining or defining the boundaries of interconnected process systems. While precedent suggests a single process is that which is piped together, recently there have been changes, or at least discussion?
Adam Kern: Determining boundaries of a covered process has always been a difficult subject, but in general between interpretation letters from OSHA and various court cases that have happened over the years (see Akzo-Nobel and MEER decision, for example), the general rule of thumb was that anything that has hazardous chemicals in it and/or has interconnected piping to that process is covered as a single process.
In 2020 a court case was decided for a PSM citation against Wynnewood Refining Co. that may have some drastic implications for extending the PSM boundary. At Wynnewood, a boiler that supplied steam as a utility to a covered process, was not considered as part of the process by the company. This was in line with previous interpretations for boundaries, but after an incident involving the boiler, OSHA issued a citation which implied that the boiler should have been part of the covered process, and that was upheld through several appeals through several courts.
The implication is that companies need to reevaluate their established process boundaries because utilities and even nearby processes may need to be included. I would probably start by widening the scope of PHAs to include more systems and assume that they can affect a covered process until you definitely prove that they cannot. Many companies already do this, but instead of officially including these systems as part of the covered process, they opt to “treat it like a covered process.”
The only problem with this is that they don’t always document as thoroughly around these systems, or they do a “PSM Light” version of managing it. This tends to lead to these systems not being taken seriously and to unforeseen accidents.
The Wynnewood decision may force companies to give utility systems and nearby equipment just as much weight as the process containing the hazardous chemical going forward.
Question: Many readers may know already the answer to this in their areas, but what are some of the most important documents to review in terms of regulation and standards when it comes to process safety?
Adam Kern: Obviously starting with the language of the PSM (29 CFR 1910.119) and risk management plan (40 CFR 68) standards would be first. After that, I would search interpretations letters from OSHA or EPA, especially as they relate to the specific chemicals you are dealing with. Beyond that, there are several publications by the CCPS on specific elements of PSM that can be quite useful for either implementing or improving performance for that element.
Question: What’s the biggest change in process safety best practices since you’ve been involved?
Adam Kern: The biggest single change relates to hot work permits. The PSM standard has a specific section for hot work which more or less requires you to follow the existing OSHA Hot Work standard 29 CFR 1910.252(a), but because it is in the PSM standard, that causes some documentation retention requirements that you otherwise might not need for other hot work. Rather than just documenting hot work, more and more I see “safe work” permits being used by companies which include other permits, like confined space or open floor/wall openings, and even JSAs (job safety analyses) sometimes to make sure that before beginning any work you have a convenient way to consider a wide variety of important safety items. This is an excellent way to ensure compliance while integrating it into your business in a way that makes sense.
Question: Are supply chain disruption or skilled labor shortages impacting process safety?
Adam Kern: I don’t think we’ve seen major effects of the recent supply chain and labor issues as of yet, but I certainly think we could see them in the near future. Skilled labor in certain areas has been an increasing problem, as there have been a declining number of highly trained welders, electricians, etc. over several years, but in general companies have been able to compensate by planning ahead or a willingness to pay more for a particular skill.
The recent supply chain and labor shortages will likely result in companies taking a leaner approach to mechanical integrity, which may well lead to some issues. The realities of running a business have to be considered over the short term, but I’d encourage any company to closely monitor any changes made in the last few years to ensure over the long term, as the world opens back up, it was a safe decision.
Matrix Technologies is a CFE Media and Technology content partner.