Murphy,
I never figured you for a chicken little kind of guy. I worked for the Bureau of Explosives for 5 years and SP Hazardous Material Control Department for 8, so propane in tank cars is something I KNOW about.
Propane is a straight chain hydrocarbon as Beauleau described. It can be liquefied at relatively low pressures, how low depends on the temperature, and there are vapor pressure/temperature charts that let you determine pressure or temperature if you know one but not the other. About 90% of the tank is liquid and 10% vapor in round numbers. It takes a page of calculations to figure out how full to fill each car based on the temperature of the product being loaded. If the lading is 55 degrees F, then pressure is 100 PSI gauge. At a temperature of 115 F, pressure is 220 PSI gauge.
Propane was first shipped in ICC specification 105A300W tank cars, which are insulated cars. In the late 1950's the 112A340W uninsulated car was introduced. The top 2/3 was required to be painted white. The general rule about safety valves is that they start to discharge at 75% of the tank test pressure, that is 225# for the 300 pound hydrotest insulated car, but for the 112 cars the specification allowed STD of 82.5%, a 10% bump, which is 280.5 pounds for the 112A340W car. Since the cars are/were not insulated, the reference temperature was 115 degrees. Since pressure at 115 degrees will be 220 pounds there is a 60.5 degree spread between maximum expected pressure in transportation and the start to discharge pressure. All is well.
You asked about leaks. Leaks will always involve the fittings. A loose plug at liquid or vapor or slip tube, or a loose valve attachment to the cover plate for example. Most of these will be vapor and so minor that they can not be detected from the ground and vapors will be rapidly dispersed and thus present no threat to anybody.
Liquid leaks are more worrysome since one volume of liquid will convert to 280 volumes of vapor. If you see frost running down the side of any liquefied gas car you are seeing a liquid leak as the liquid sucks heat out of whatever it touches so it can boil off.
I had a liquid leak from the safety valve in Millersburg Oregon one night. This is not supposed to happen. Remember the page of calculations? This is not diffucult to diagnose, the car is absolutely full of liquid. Solution, call up the local propane dealer and give him a delivery truck's worth of propane for his trouble, which is what we did. The next morning I a got a call from the shipper up in the Great White North around Edmonton. He was irritatated that I gave his gas away. I told him he gave it away when he overloaded the car in violation of the Hazardous Material Regulations. End of discussion.
I never liked or used the word BLEVE, which some fire service author make up to scare the fire laddies. He accomplished that purpose, but at the cost of confusing what was going on and thus making predictions about what was likely to happen, and what to do more difficult. I always called it flame induced violent rupture, and illustrated it with what I called the teakettle effect. Most people have the experience of putting water in a teakettle, putting it on the burner, and getting a plume of steam. No problem UNLESS you boil all the water out and you melt the teakettle.
In the context of Propane cars, the problem is heat input or flame impingement on the portion of the tank that does NOT have liquid behind it. The Kingman incident is the poster child for this. Somehow the guy unloading the car broke a liquid hose from the pipe that extends through the manway bonnet. Odds are he did not open the valve all the way since there is an excess flow valve in the tank car's liquid line that should have functioned and shut off the flow. Somehow someone managed to set the vapor from the liquid on fire. That created a two inch diameter blow torch parallel to and about 8-12 inches above the uninsulated tank shell which of course had no liquid behind it. The flame heated the tank, which has a designed burst pressure of about 850 PSI, to the point that the tank failed due to excess heat dropping the tensile stength of the steel to something below the internal pressure being maintained by the safety valve, which was about 280-300 pounds. At that point the tank failed. The liquid then flashed to vapor and burned in a few seconds making a large fireball above the tank car. It was spectacular but not an explosion.
Mac McCulloch
