Pipeline residue handling

[oldfieldguy]

Forgive me for a simple question, but I'm and electrical guy looking at a little faux pas we had out on a natural gas pipeline.

During a pigging run, residue from the pigging operation was pumped into a portable "frac tank", a large mostly rectangular, closed top vessel. Somewhere in the course of the operation, we had an ignition inside the frac tank, resulting in the top splitting and the sides bulging out.

The pattern of the bulge seems to corroborate the 'explosion' idea, except there was no subsequent fire.

There is a pressure relief on the tank and it appeared to be in working condition.

I have NO idea what manner of substances came out of the line during pigging, but I would suspect a variety of hydrocarbons ranging from stuff that is barely liquid at ambient temperature and pressure, right on down to heavier compounds.

My question: The frac tank was NOT inerted. They began pumping into the tank while it was filled (empty) with ambient air. MY question was "is this the way it's always one?" and was told "yes". So question to the group: Should that tank have had an inert atomosphere? Anybody who has experience in these things?

thanks

old field guy

 

[ash9144]

Not to jump on an obvious answer but since you had an explosion the answer to inerting would be yes it does need that.

In practice no it isn't normally done. I only recall seeing it done on one frac tank that was being used for benzene storage.

One item other item which is a key thing to ensure is that the boxes are grounded. That could have been your ignition source.

Please share your findings as I think many of us could learn from this incident.

 

[oldfieldguy]

ash--

That grounding thing was how I was sucked into the investigation. It appears that grounding and equipotential bonding was properly installed and adequate.

The inerting question arose because I spent quite a bit of time in refineries and petrochemical plants and it seemed they didn't do anything with hydrocarbons of this nature without inerting, but again, I'm electrical and only peripherally involved, so I didn't pay attention to details.

It's just that in this day and age "we've ALWAYS done it that way" is a poor excuse.

old field guy

 

[zdas04]

Inerting the tank is a very expensive undertaking and is not necessary.

If I was on the investigation team I would look for a grounding strap. High velocity fluids moving over a pipe surface will develop a certain amount of static. Nothing will prevent this. If the tank is properly grounded, then the charge will dissipate to earth before it gets big enough to arc. If the tank is not adequately grounded (and as a Sparky, you should be able to assess this better than most of us Flange Heads) then you can build up enough charge to arc and blow the tank up. This happens all the time.

An explosion without a fire is pretty common. There is a limited amount of air in the tank. When you've created an explosive mixture and ignited it with a spark, the unused air is usually not adequate for further combustion (i.e., the mixture is now richer than the UEL) and the fire never starts.

Inerting the tank is not a trivial undertaking. There usually isn't a reasonable flow path to do a clearing purge, so you are limited to dilution. If you start with an "empty" tank full of air you would have to raise the pressure to around 7 psig to get to a mixture that would be assured of not supporting combustion. At 7 psig the frac tank would come apart so you have to do it in stages and quality control of the stages is a difficult problem. Basically you would have to pressurize the tank to a safe number (say 1 psig), blow it down to half that number, and pressurize it back to the safe number. Repeat this until you think that you are safe, gas monitors are not a lot of help in a tank problem because the flow path bypasses most of the tank and homogenization of the gas is a time-dependent function.

David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

"Life is nature's way of preserving meat" The Master on Dr. Who

 

[zdas04]

OldFieldGuy,
I think we must have hit enter at the same time.

You need to look for an ignition source. I would re-examine the grounding because static is the only ignition source that is reasonable for most of these sorts of explosions. If there is another ignition source then you need to find it and try to render it safe. Inerting the tank is probably not a reasonable solution.

David

 

[Latexman]

If the residue was non-conductive enough and the entering velocity high enough, static could be generated and build to the point where a healthy spark finally occurred. Inerting is the classical solution. Controlling the inlet velocity is a poor second choice, imo.

Good luck,
Latexman

 

[oldfieldguy]

zdas--

Thanks! Very informative and explains a lot of what is going through my mind.

We actually had several frac tanks in the area, all bonded together and to a grounding rod "triad". Other equipment in the area was also grounded.

What about conductive or "anti-static" hoses? I don't know if we specified those or if they were used.

I am familiar with the phenomenon of charges building as a non-conductor (the hydrocarbon stream) moves from one conductor to another. Our equipotential bonding is supposed to take care of that, and I personally inspected the connections after the fact, and they appear to be correct.

old field guy

 

[Latexman]

There are documented cases of well grounded containers with flowing non-conductive flammables having explosions and fires. IIRC, benzene and heptane are notorious for this. It seems the static charge buildup can exist in the center of the volume and not get dissipated through the liquid to the walls. Then an arc occurs, maybe the liquid level approaches an internal part of the container, and . . . BOOM! Inerting and/or controlling velocities has eliminated the concern. I prefer inerting, but I live in a chemical plant where nirogen is cheap and plentiful, not "out yonder" where the gas is collected.

Good luck,
Latexman

 

[zdas04]

I've actually had more problems with static in hoses than in fixed pipe. The ones with a woven metal sheath seem to do better, but still not great.

This type of explosion was really common in the early days of fiberglass tanks. You could open a tank that you thought was gounded and find solid material dancing on the surface above the liquid. The problem was usually that the liquid was non-conductive (a good thing for corrosion issues, not so good for explosion issues) and the ground was limited to a short rod in the bottom of the tank. Some tank manufacturers fixed this problem with wires bonded to the inside of the tank to carry to a common ground. This worked well.

Could it be that your frac tanks have inconsistent electrical conductivity properties that can allow an arc to jump a non-conductive section? Maybe a weld joint that is non-conductive that allows a charge to build up above the weld and arc to the grounded section below?

I'm focusing on static because it is easily fixable. Other ignition sources are a lot harder to find and/or repair.

David

 

[Latexman]

A good reference is "Avoiding Static Ignition Hazards in Chemical Operations". I've worked with the author, Larry Britton, when he worked for Union Carbide. He knows his stuff! I think he consults now.

Good luck,
Latexman

 

[ash9144]

From the sounds of it the tank was well grounded.

I would look at the specifics of what was going on at the time. Hoses being connected, temperatures, opening lids, etc.

Also you mention you had an explosion based on the tank distortions. Are you sure this was an explosion and not just an overpressure event?

As for frac tank inerting I don't think it is very difficult to do for these type applications. Nitrogen hose sweeping out to atmosphere and your O2 concentration will come down fairly quickly. Once you start filling with hydrocarbon your nitrogen will be displaced with the HC vapors.

 

[zdas04]

I've always heard that the definition of "easy" is "someone else has to do it". I was on an accident investigation team once where they were blowing nitrogen into a frac tank to "inert" it. Problem was that the nitrogen channeled and there was a pocket of air that didn't get swept. I've been nervous about "just inerting" ever since.

It is pretty easy to tell the difference between an overpressure event and an explosion--overpressure event's don't turn the slime on the tank walls to soot.

David

 

[Latexman]

I think inerting in vessel designed to be inerted in a chemical plant, where you have a huge liquid reservoir and miles of distribution pipe, is easy. Out in a gas field with cylinders hooked to a frac tank is another story. Main reason being it takes time, no dead spaces, and volume of N2 to get down to non-flammable. And let's not forget $.

Good luck,
Latexman

 

[Compositepro]

Charge build-up is always an equilibrium between rate of charge generation versus rate of charge decay. Charge is generated when non-conductive fluid flows though pipes - even grounded metal pipes. The mechanism is separation of fluid molecules from the pipe surface. The fluid molecules have a greater or lesser hold on electrons than the pipe material so the fluid will accumulate positive or negative charge. In a non-conductive fluid the charge is attached to molecules and can only move with the molecule. Charges are attracted to opposite charges and repelled by like charges. So charged molecules in a pipe or tank will be pushed to the outside by repulsion of other charges and they will be attracted to grounded surfaces.

The charge in a grounded pipe will dissipate pretty quickly when flow stops, due to this effect. If there is a non-conductive "slime" on the walls of a tank, it can act the same as a non-conductive rubber liner and create a capacitor(Leyden Jar).

Using this understanding can lead to a number of different solutions. One common approach is to fill tanks though dips tubes so the fluid enters the tank below the liquid surface where it slows and then the charge can reattach to the pipe from the outside. This would tend to slime-up as well, but it can be removed for cleaning. Or fill the tank slowly. It may take less time than inerting.

 

[oldfieldguy]

latexman-

Your comment lines up with my theory. Unfortunately in my case it was simply a "I wonder if this could happen" thing.

zdas-

All the hoses were connected and the project was in the "stand around and wait" stage as far as I can tell. One of my fist comments in the investigation was about the sparking ability of ferrous metals as an ignition source. We made extensive use of hoses in the project. I do not know if they were conductive or not.

Compositepro--

I used the Van de Graaff generator as an example of static charge built up by moving a non-conductor from one conductive terminal to another. I was figuring that our equipotential bonding of metal components would drain off this charge, but "latexman" supports a nagging idea from the back of my mind.

To all--

This has been a very educational experience for me. You guys are great!

old field guy

 

[HEC]

A couple of things to look at, was bonding installed across all flange and other connectinos on the pipework and hoses. I know real basic and you have already said it has and is adequate. Second item, where is the inlet to the tank, and what sort of velocities are generated. If the inlet is near the top of the tank is there a long free fall to the fluid already in the tank. If below the liquid surface are high velocities present that can create turbulence?

Mark Hutton

 

[oldfieldguy]

Mark--

No, we don't bond across flanges. This is a temporary lash-up for one of these pigging operations. Part of the piping is made up of hammer unions, part with rubber hoses with various metallic fittings at each end, and a little part with bolted flanges.

About the inlet and internal flows, the inlet is at the top of the tank, so if there's a significant flow into an empty tank, it has quite a drop. I am under the impression, however, that eh tank was more than half full, so the drop should have been less than four or five feet, adding of course, any projection due to stream velocity.

I can't see where with the size of the tank (8'x8'x40') and the entrance line size (around four inches) that we'd have very much velocity at the liquid surface, but I could be incorrect in my visualizations.

old field guy

 

[HEC]

old field guy

Fluid velocities above 1m/sec are considered significant. It is better to direct the inlet flow against the side of the tank to reduce the free fall issues.

Mark Hutton