Fire at the Filling Station

Static electricity can be a source of ignition. Static ignition has started fires involving people filling their vehicle tanks with gasoline. Static charge buildup is generated when liquids move in contact with other materials, such as when gasoline flows through a hose and nozzle. The vehicle in this photograph experienced a fire caused by static ignition.

Vehicle service station pumps usually have warning signs posted about the danger of static ignition. The signs have instructions on how to reduce the chance of it happening. The gasoline nozzles also have warnings that can be read while you are filling the tank. One warning says to keep your hand on the nozzle while filling the tank. Do not set the nozzle on automatic fill and leave it to go get something out of the car, or to go into the store, or whatever.

This is because while you are sliding across the car seat or walking around, you can build up a static charge that is different from the one that is building up on the nozzle. Then, when you return to the nozzle and touch it, if the difference between the static charge on the nozzle and the one on you is great enough, a static discharge will occur. This creates an electric arc between your hand and the nozzle. If the electric arc has enough energy, and if other conditions are right, the gasoline fumes will ignite.

Also, if you are filling a portable gasoline can, place it on the ground first. If you fill the can without first removing it from your truck bed or car trunk and placing it on the ground, you are courting the possibility of static ignition.

This is a photograph of a warning on a gasoline pump at a service station near my home.

This is a photograph of the warning on the gasoline nozzle:

I Like Ockham's Razor

Ockham’s Razor is a principle espoused by William of Ockham in the fifteenth century that translates as "entities should not be multiplied unnecessarily". In other words, if you have two theories that are both consistent with the observed facts, then you should use the simplest theory, unless and until evidence to the contrary is revealed.

For any given set of facts there are an infinite number of theories that could explain them. For example, if you have a graph with four points in a line, then the simplest theory that explains them is a linear relationship. You could draw an infinite number of different curves that all pass through the four points, and there is no proof that a straight line is the right one. However, it is the simplest possible solution. Therefore, you might as well use it until someone comes along with a point that is not on the line.

The Day the Slop Oil Tank Blew Up

I’ll bet there have been times when you heard about some big explosion at a refinery or chemical plant or whatever, and you wondered what the cause was. Or you’ve heard some TV announcer say something like, “Investigators at the scene are trying to determine the cause…” But have you ever considered the possibility that it might have resulted from a chain of events, not just one, and that if even one of those events had not occurred, the explosion would not have happened?

Well, I have. I learned a long time ago that explosions sometimes happen exactly like that. The ones I’m going to tell you about occurred several years ago. A refinery slop oil tank blew up, releasing boiling hydrocarbon liquid and volatile gases. They formed a huge explosive cloud that exploded several minutes later

Refineries have slop oil from spills, off-test products and whatnot. It is captured in the wastewater treatment unit by a separator that operates on the simple principle that oil floats on water. The oil is decanted off from the separator and heated to 250-300 degrees Fahrenheit to lower its viscosity, making it easier to pump. They pump it to a slop oil tank, and from there they feed it back into the refining process.

You might be wondering why a slop oil tank would blow up. I mean, it’s an uncommon thing to happen. You might be wondering why boiling liquid was involved. You might even begin to suspect that something other than slop oil was in the tank. You would be right. It turns out that the refinery was using the slop oil tanks temporarily to store highly volatile hydrocarbon liquids produced by the cat cracker.

The catalytic cracking unit is a major processing plant in an oil refinery. It is commonly referred to as the cat cracker, or the cat unit. It breaks up long-chain petroleum molecules and converts them into propane, butane, and a variety of intermediate liquid products used to make gasoline, jet fuel, and heating oil.

Propylene and butylene are cat unit products. These compounds are called olefins (which is merely a generic label for organic compounds with one or more double carbon-carbon bonds). Typically, these particular olefins are used as feedstock for the alkylation unit. The “alky” plant, as it is typically called, chemically combines the olefins with butane to make iso-octane, which is a key component of gasoline.

The alky plant at this refinery used hydrofluoric acid as a catalyst. (Some alky plants use sulfuric acid.) For a couple of days prior to the explosions, the acid pump seals kept failing and shutting the plant down. The acid was disintegrating the pump seals and spewing out all over the place.

Now think about what’s going on here – the pump seals fail, and the plant shuts down. They replace the seals with the same type, and start up the plant again. The seals fail again, and the plant shuts down again. They replace the seals again with the same type, and start up the plant again. And they do this repeatedly, apparently expecting different results each time they do it. Duh?

Well, surprise -- it later turned out that the seals were no good for acid service. Someone purchased the wrong type. This could have been a purchasing error. Or it could have been that these seals were cheaper than the right kind. Anyway, this was the first event in the chain. The second was when the people installing the seals did not catch the error.

With the alky plant out of commission, there was no place to put the cat unit olefins. The refinery manager was not about to shut down the cat unit. Instead, he decided to put the olefins into railroad tank cars. This worked out kind of okay until he ran out of tank cars. Then he decided to pump out the slop oil tanks, and put the olefins there for the time being. He also had a written order entered into the wastewater treatment plant manager’s log, saying not to pump any slop oil to the slop oil tanks until further notice.

Now, slop oil tanks are vented to the atmosphere. At atmospheric pressure, propylene and butylene boil at temperatures well below zero degrees Fahrenheit. Consequently, the olefins in the slop oil tanks were boiling, and the vapors were being vented. Moisture in the air froze on the outer surface of the tanks below the liquid level. Ice coated the exterior surface of the tank from the ground to the liquid level.

Meanwhile, down at the wastewater treatment plant, things were getting dicey. Slop oil piling up in the separator would soon start overflowing into the river. The plant manager wasn’t about to get fired for letting that happen and getting hit with huge cleanup costs and EPA fines for killing the fish, not to mention for losing product. So, he ordered one of his men to drive over to the tank farm and look at the ice on the slop oil tanks to see if there was any space left in either of them (there were two slop oil tanks).

Take note here, that the wastewater treatment plant manager has no clue as to what might happen if he pumps 250-300 degree slop oil in with the olefins. He knows the ice forms on the tank because the olefins in it are boiling, but he doesn’t equate that knowledge with anything other than how full the tanks are. He’s thinking only in terms of how much room there might be left in the tanks.

Anyway, when the wastewater treatment man got to the slop oil tanks, he radioed back to his plant manager that there was about four feet of room still left in the top of one of them. So, the wastewater treatment plant manager – with complete disregard for the refinery manager’s written order – started pumping the hot slop oil to the slop oil tanks at about 200 gallons per minute.

Now, any given volume of propylene expands approximately 300 times when it boils into gas at atmospheric pressure, and butylene expands about 200 times. When the hot slop oil reached the tank, it quickly vaporized huge amounts of the olefins. The tank vents were woefully undersized to accommodate the enormous volume of gas suddenly being created inside the tank. The vents were designed to handle normal tank “breathing,” not something like this. The resulting sudden overpressure blew the roof off, releasing an enormous volume of boiling hydrocarbon liquid and volatile gases.

You probably know that the range of concentrations of hydrocarbon gases in air that form an explosive mixture is typically fairly narrow. For example, gasoline vapor will not ignite unless there is at least about 1.4% gasoline vapor in air. On the high end, gasoline vapor will not ignite if there is more than about 7.6% gasoline vapor in air. So, the “explosive envelope” for gasoline is about 1.4-7.6% in air. Corresponding values for propylene are 2.0% and 11.1%.

Thus, for the gas to be explosive, it must be diluted tenfold or more with air. To put this into perspective, suppose that one cubic foot of liquid propylene forms 300 cubic feet of propylene gas. This is diluted with air to form between 2,700 and 15,000 cubic feet of explosive gas-air mixture, depending on the dilution factor (300/0.111 = 2,727, and 300/0.02 = 15.000).

So, the olefin vapor cloud floated out over the refinery, with air diluting it as it went. By the time the cloud contained enough air to make it explosive, it was enormous, and it was searching for a competent source of ignition. It found it about 300 yards away, where three men were working with welding equipment, and a tremendous explosion occurred.

All four men died -- the wastewater treatment man at the slop oil tank and the three men in the welding crew. Several million dollars worth of refinery equipment was destroyed.

These explosions clearly illustrate what I said in the first place; namely, that explosions may result from a chain of events instead of just one. Neither one of these explosions would have occurred if any one of the following things had not:

• The purchase of alky plant pump seals that were not appropriate for acid service;
• The failure to catch the error when the seals were first installed;
• Repeatedly installing the same type seals while expecting different results;
• The decision to store olefins in the slop oil tanks; and,
• The decision to disregard a special written order and pump hot slop oil into the tanks while they contained olefins.