Earlier this week I posted about how a PHA should document its consideration of the engineering controls and administrative controls failing. And I said back in the 2013 posting; this little requirement is intended to make facilities dig past a single layer of protection and maybe even identify a lack of engineering and/or administrative controls for a process deviation. But let’s be clear, a Process Hazard(s) Analysis is a SINGLE TOOL we use to identify “hazards” of our process and the safeguards we have in place to PREVENT the scenario from initiating, PROTECTING assets once the event has occurred, and then MITIGATING the consequences of the event. In my post earlier this week, my layers began at the OPERATIONAL LEVEL of process control. What do I mean by this?
Our true FIRST layer of protection against a Loss of Primary Containment (LOPC) event is the actual PRIMARY CONTAINMENT. Using the same scenario of HIGH PRESSURE in a vessel, as I did in my earlier posting, we should ALREADY HAVE engineered and analyzed the consequences of deviation from the DESIGN when we were establishing our Process Safety Information (PSI)…
1910.119(d)(2) Information pertaining to the technology of the process.
…
1910.119(d)(2)(i)(D) Safe upper and lower limits for such items as temperatures, pressures, flows or compositions; and,
1910.119(d)(2)(i)(E) An evaluation of the consequences of deviations, including those affecting the safety and health of employees.
We know this must be done BEFORE we even begin our PHA, because of…
1910.119(d) Process safety information. In accordance with the schedule set forth in paragraph (e)(1) of this section, the employer shall complete a compilation of written process safety information before conducting any process hazard analysis required by the standard.
So the reason I do not look at the pressure vessel catastrophically failing due to HIGH PRESSURE during the PHA is because the facility is supposed to have ALREADY done this analysis when they were putting their PSI together (e.g. .119(d)(2)(i)(D) and .119(d)(2)(i)(E)). But the vessel itself is ABSOLUTE KEY to process safety, and during the PSI compilation, we should have been challenging our process design even BEFORE this vessel was ordered!
For example,
we told the engineer this vessel would be part of an ammonia refrigeration process. And based on our process needs the vessel needs to be AT LEAST a 10,000-gallon vessel, as we need AT MOST 8,500 gallons safe storage capacity in this vessel (e.g., assuming we use 85% as our SAFE UPPER LEVEL). The engineer would then go to the LIST OF RAGAGEPs employed in the design, construction/installation, Operation, and Maintenance of this process and see that his/her chosen RAGAGEP has established this vessel must have at least a Maximum Allowable Working Pressure (MAWP) of 250 psig. However, that is the ONLY established “minimum” this RAGAGEP requires, and the left is UP TO the engineer to establish and design SAFETY FACTORS he/she (or maybe the business has established) is comfortable with. The chosen RAGAGEP does make mention of some other critical process parameters such as “temperature” by stating:
Equipment shall be designed to operate within the full range of temperatures associated with the system design and for the full range of ambient temperatures to which equipment will be exposed at the installation location. The manufacturer shall provide operating temperature limit information.
But let’s get back to the vessel and begin asking questions:
Since this vessel has to have a minimum MAWP of 250 psig, we already know this vessel MUST be a “pressure vessel.” This brings in ANOTHER RAGAGEP, and since this will be an UNFIRED vessel, we would use ASME VIII construction criteria and ensure the. Thus is fabricated AND STAMPED by a certified manufacturer for the design parameters we require.
Since the MINIMUM pressure this vessel must meet is 250 psig, is that adequate for our process conditions AND atmospheric conditions? Where will this tank be sited? Inside? Outside? If inside, what is the temperature(s) it could see? If outside, what is the temperature(s) it could see? What if our process will be outside in Phoenix, AZ where our vessel could see ambient conditions of 115°F for days? Using Anhydrous Ammonia’s VP curve we see that at 115°F the pressure inside this vessel would be 251 psig; in other words, just from ambient temperature conditions, we would have a LOSS OF PRIMARY CONTAINMENT! Sound stupid – it should, but it happens!!! So, we would probably want to go up to a MAWP of at least 300, and I would strongly recommend 325°F as I like a more substantial “safety factor/margin.” But so far, we have only looked at MAWP and MAWP is tied to the vessel’s MAWT.
The Maximum Allowable Working Temperature (MAWT) is tied to the MAWP, such that as the temperature of the metal increases, its strength begins to decrease. Hence why we see the MAWP displayed as:
MAWP = 300 psig @ XXXF
That XXXF is another CRITICAL path for our PRIMARY CONTAINMENT. If this vessel is merely a storage vessel sitting outside/inside and has no type of heating or cooling to it, then we would start off by considering is the AMBIENT temperatures it would be exposed to. Such as how hot could the room this tank will sit in get to? We should ALWAYS consider the loss of HVAC if the room is climate controlled, as I have seen too many times an uninsulated vessel sitting full for days on end during a turn-around and the room gets way outside its normal temperature range, and this caused an LOPC event. I would also mention here that this XXXF has NOTHING to do with our “fire scenario” when we do our RV design and sizing analysis, that is an entirely different exercise we can discuss another time. In the sake of time, I will direct you to read “A pressure vessel service change (Lesson Learned)” to learn more about how the MAWT of a vessel MUST be considered in the service and location of a pressure vessel.
The next CRITICAL path for our PRIMARY CONTAINMENT is the Minimum Design Metal Temperature (MDMT); much like our metal loses strength when it gets too HOT, that same metal will become too brittle when it gets too COLD, thus the vessel MUST have an established MDMT based on our process and ambient temperature conditions. Most ASME pressure vessel will have a standard -20°F as their MDMT (i.e. if your nameplate and/or U-1 form is blank for the MDMT data your vessel most likely has an MDMT of -20F); however, if our vessel will be sitting in a location where it could see sustained temperatures COLDER than -20F (e.g., Alaska/Canada for example) then we will most likely need a SPECIAL MDMT for just the ambient conditions. If our process can create conditions where this vessel would see temperatures less than -20F, then we would need to ensure the vessel has the correct MDMT (including proper safety factors/margins) for our process conditions as well.
So as the engineer(s) is going through this design process, they are PUTTING PROCESS SAFETY FIRST, but also trying to design the process in an economical manner. Sure, they could just use pressure vessels with a MAWP of 1,000 psi @ 750°F and a MDMT of -100°F, but God only knows what just the price of the pressure vessels alone would be! So those ranges are way off the charts and we begin to ask ourselves what if (no pun intended!) we did have a MAWP of 250 psig and our plant was in Phoenix AZ – will that work? We determined it would not using just the ambient conditions and the properties of our hazardous material. So how about 275 psig? 300 psig? 325 psig? In each consideration we are asking ourselves:
What does the MAWP safe upper limit for pressure need to be so as to NOT have a LOPC event that would affect the safety and health of employees? (this question should look VERY familiar and eerily close to some PSM/RMP language!)
The same evaluation/analysis would occur for the MAWT and MDMT of these vessels. So when it is all done, the process will have established (for its vessels):
1910.119(d)(2)(i)(D) Safe upper and lower limits
1910.119(d)(2)(i)(E) consequences of deviations affecting the safety and health of employees
This is DONE as PART of the PSI work and well BEFORE we even attempt to perform a PHA! So the reason my example of safeguards earlier this week started with a PSI indicator notifying an operator of a pressure deviation AND not with the vessel has a MAWP of 300, is because I have already evaluated/analyzed my vessel and its design capabilities for my PROCESS and AMBIENT conditions. In my PHA I want to focus on what is going on inside (and outside) of the vessel that is causing the HIGH-PRESSURE deviation and what engineering and administrative controls I have in place to PREVENT the process from entering the deviated state, PROTECT my assets when it does enter that dangerous state, and how I am going to MITIGATE the consequences of the event. But I know one thing… my PRIMARY CONTAINMENT is solid as I have ALREADY established its safe upper and lower limits AND evaluated the consequences of deviations from these limits that may affect the safety and health of employees.
Lastly, I want to make it CRYSTAL CLEAR… the PSI exercise I have discussed here is NEVER to be done in a “vacuum” or “silo” – meaning NO ONE PERSON should have the ability to establish these safe upper and lower limits on their own. In my world, we called this exercise a PRE-CONSTRUCTION PHA, and these were done with a TEAM of professionals who had the skills and experience with the process and the HHC/EHS such that we knew what “extreme” conditions we needed to design to. Remember, much like OSHA/EPA standards, RAGAGEPs often establish MINIMUMS as they were broadly written. It is OUR RESPONSIBILITY to take these established minima and work beyond them as needed based on our process conditions. If we get to a PHA scenario and we find that a “process deviation” can take us outside our PRIMARY CONTAINMENT capabilities we are looking at a MAJOR issue – one that may require the process to be run DIFFERENTLY than intended UNTIL such time that the design flaw can be repaired. This can have a crippling effect on a business, as well as increase operational risks; however, we should NEVER experience this when we have COMPILED our PSI properly.