Quash Commissioning Quandaries

This Month’s Puzzler

Steam was used to sterilize and then pressure test a new 12-ft-dia. storage tank for our reactor train. Unfortunately, the tank shell ruptured. It was rated for 45 psig, not vacuum. It didn’t have a U stamp; our project engineering group decided not to spend the $3,500 required. The old tank is in the bone yard. There isn’t anything wrong with it, except that is about 25% too small and was de-rated to 20 psig; in fact, its R stamp is current and it’s rated for full vacuum.

There were a few things I observed after investigating for a day: 1) maintenance staff conducted the work because the people in operations were tied up in re-training during the downtime (the week between Christmas and New Year’s); the instrument tech and the youngest mechanic on our crew admitted they’d never done this type of work before; 2) the project engineer was busy on other projects while the superintendent was away on vacation and out of touch; 3) the top manway cover that was supposed to be closed during the pressure test but open during the steam sterilization was bolted tightly (sterilization follows pressure testing); 4) a ½-in. flange on a gasketed vent line was loosened to act as a vent during sterilization — the crew was given this advice over the phone by a production foreman taking a break from skiing; and 5) a fire hose was used to fill the tank for the pressure test, which seems extreme to me.

What can we do to avoid this problem in the future? What can be done now to get us up and operating?

Revamp Commissioning

The troubleshooting request fails to mention the pressure/vacuum relief system configuration and the vessel vacuum rating, nor specify explicitly whether this is an overpressure rupture or a tear from vacuum collapse. The problem statement implies it could have been either. Regardless, formal constraints for under- and over-pressure are needed during the commissioning/startup phase. In other words, the steam supply crosstie to the vessel and the method to limit flow/pressure would be established by engineering calculation and the connections, depressurization, padding and vents sized and specified in the commissioning/startup procedures, so that the relief system is not overwhelmed, unsafe conditions are prevented and commissioning steps are of an acceptable duration. Even if not required by process safety management, consider a rigorous safety review (process hazard analysis) of the commissioning procedures following a standard methodology.

Having a code stamp is no guarantee that the vessel or relief was properly specified for the commissioning conditions. The code stamp itself in a non-code state would only be a validation of quality for the specified construction. So far as the information available, one cannot say whether fabrication itself was at fault or not. One might require including commissioning conditions on the vessel specification sheets issued for bid/fabrication, so that the vessel and/or relief are appropriately designed. As a side note, although the pressure vessel is rated for 45 psig, it is unlikely to withstand vacuum at a 12-ft diameter unless explicitly designed for vacuum. (Liquid drainage and steam collapse are well known to generate vacuum and should have been included in the relief scenario evaluation if not a full vacuum-rated vessel.)

Restarting with the old vessel may be feasible but all the same issues are present and must be addressed, unless the derating calculations confirmed it still has a full vacuum rating.
Gary Holleran, senior process engineer (retired)
BASF, Beaumont, Texas

Fire The CEO!

I couldn’t believe what I was reading in the Puzzler. This company needs psychological help and is a disgrace to the chemical industry and chemical engineering profession.
I can bet you that even a ten-year old would fire the chief executive officer of this company. I am being honest —the company needs an honest opinion; they are jeopardizing lives.
Girish Malhotra, president
EPCOT International, Pepper Pike, Ohio

Tackle Several Issues

One of the chronic problems companies face is keeping personnel up-to-date and happy. The mistake here can’t really be blamed on two people caught holding the dirty end of the stick. Nor can it be blamed on the foreman: people need vacations. When I took an informal poll of about ten production engineers, none would recommend production work to their children — regardless of what it paid.

No, the blame lies with the planner. It’s a tough job because the planner has to be ready to substitute qualified people in case someone is sick. Some blame goes to the project engineer for not being present, given the limited skill of the instrument tech and young mechanic — knowing your people is part of good leadership. However, it may be that the project engineer didn’t know any better either. Other people who should be in line for a chewing-out include the safety manager, the superintendent, perhaps the trainer and probably the maintenance manager.

The post-mortem review of this accident should include an analysis of training requirements. Everyone onsite who might be called out to do a pressure test should know how to do one. Don’t forget the managers; they should be able to teach the class in their sleep.

So, what went wrong? Steam condensed, causing a vacuum in a vessel that wasn’t rated for full vacuum or there was a mistake in the vessel design; that’s one reason why you pay for the U stamp. Somehow the directions were mixed up, so that steam was used for the pressure test and the sterilization. The manway should have been open for the sterilization to prevent condensing steam from collapsing the vessel. The underlying problem is that the mechanic, the instrument tech and the foreman on the phone apparently didn’t understand what happens when steam condenses.

As for using a fire hose to fill the tank for a pressure test, that seems fine to me. I’ve seen it before. Filling it and then pumping it up to 1.3 times the rated pressure, 58.5 psig, (changed from 1.5 as of ASME 2010) with a pump is a common method. Why wasn’t this done? Perhaps because it seemed easier and faster to use steam for the test and for sterilizing the tank while still closed. This an improper approach for sterilization because you want the steam to provide a steady heat flux at an autoclave temperature. Condensing steam doesn’t do that.

What can be done now to get the plant up and operating? Well, the obvious pathway is to re-use the old vessel in your boneyard. You have two questions for operations management:

1) Can they live with 20 psig?
2) Can they work with a tank that is 20% too small?
Note that 20 psig limits the operating pressure to only 16 psig unless relief valves with seals are used, then 18 psig is allowed.

The other questions are for the project team. Will removing the tank and re-installing it de-rate its pressure rating further? How much damage was done to the tank while removing it? And how much will it cost to repair that? How will the piping mate-up against the old tank? Piping often is changed — re-routed or re-sized. And, lastly, how much will all this cost? It may take one-to-three months to repair the tank and another month to re-install it. Or it may be possible to repair the tank in-place, depending on how this affects the plant operation; this usually isn’t a good option if it requires a shutdown. Lastly, explore other ways to increase capacity while keeping the old tank for a year or so. Perhaps reactor optimization is possible but hasn’t been assessed because installing the new tank was easier and required less engineering time. Carefully consider all these ideas.
Dirk Willard, consultant
Wooster, Ohio

June’s Puzzler

I am a new graduate now working as a production engineer at an extrusion plastic manufacturer and would like to make a good start. I manage a process to produce paint pigments and would welcome some tips on optimizing the operation (Figure 1).