I recently read that the typical shell on a tank car is made up of only about 1/10" thickness of metal, and the rest just an insulation layer. Is that really typical? it would sure explain the pics of the collapsed cars that result from being sealed after a hot cleaning, or emptied without proper venting… but sitting here with a 3/32" drill bit in hand and realizing that 1/10" isn’t much more than that, it’s hard to believe they’d build those cars so frail.
That’s the shell over the insulation. Under the insulation (often about six inches thick) is the actual tank, which I’m pretty sure is a lot thicker.
It’s interesting to see these huge tanks, labeled for the transportation of LP gas, chlorine, and other such commodities with tiny square patches in this shell. I’m sure the actual tank wouldn’t be allowed to be patched like that–nor do you see such patches on uninsulated tank cars.
When Union Tank Car Company used to build tank cars in East Chicago, Indiana, one could drive (or bike!) along the west edge of the property and see unfinished tank cars, and sometimes components of others. You could see provisions made for adding the insulation over the tanks there, and the tank heads (ends) were definitely thicker than a tenth of an inch.
Ahhh, that makes more sense, perhaps I wasn’t reading carefully. [:I]
We just talked about that in my hazmat refresher course, but I can’t remember those numbers.
A little on-line searching found thicknesses of 9/16" and .603" for tank car shells, with the former also applying to the ends.
I found tank wall thicknesses from 0.4375" to 1.136".
Isn’t the force from coupler to coupler passed through the tank walls or is there some other internal beam that carries the load?
Yes - for those tank cars that have the ‘stub’ end sills, the coupler forces are indeed passed through the tank walls. And, it’s the tank itself that acts as a short, stubby beam to carry the weight of the car and contents, and as a horizontal column or tension member to carry the coupler forces. Think of the cardboard cylinder core of a standard roll of toilet paper, but loaded evenly across the ends. But those forces are surprisingly small for the size of the car.
For example, for a 10 ft. = 120 inch diameter tank, the circumference is Pi x 120 = 377 inches; if the wall is 1/2’’ = 0.50’’ thick, that’s an aggregate area of 188 square inches. For coupler ‘draft’ = tension or pull forces of 360,000 lbs., that’s an average tensile stress of 1,915 PSI. That’s only a small fraction of what the steel can take - maybe 10 %, depending on the specific steel and design criteria, etc.
To that must be added the stresses from the weight of the car and contents, plus the stress from any pressure those contents may be under, and perhaps some others that I can’t think of right now. At the moment, my intuitive reaction is that the pressure would be the biggest potential concern - the tank car is so tall and short as a beam that those stresses probably are not what control the design. But since I don’t do that for a living, I could be wrong.
- Paul North.
I was an inspector for the Bureau of Explosives in a past life and provided technical support to the AAR Tank Car Committee. Excluding “cryogenic” cars, the minimum tank shell and head thickness that I recall for steel tanks was 1/2 inch for DOT 111A100W1 cars. A W1 may or may not be insulated. If a car was insulated, the insulation was protected by a metal jacket. Your 1/10 inch sounds about right but I do not have the specifications handy to confirm or deny.
Paul - The term “Wall” is not used in tank car construction. The correct terms are shell, which in the case of a stub sill car, as you explained carrys draft forces. Each end of the shell is closed by a “head”.
Mac
I did a few Thicknes tests on Haz-mat tank Trailers in the mid 1990’s now the FMCSA has a standard we had to meet back then. SS had to be a min of 3/8in in thickness less than that the trailer had to be rplaced NO repairs allowed. Alloy trailers was 1/2 in thick. We did not use standard steel for one reason what we hauled would EAT IT like no tomorrow. Put it to you this way boss wanted to install a new fuel tank we had a set of bolts for an emergancy shut off valve fail 4 bolts on 5 thousnad gallons of acid. It fell on the yard of the terminal it dug the hole for us as it ate thru the crushed limestone rock of the parking lot. We called Hazmat in they said well you did what we would have hit it with lime nothing more to do. Boss was able to put in a 9K gallon underground tank with min digging.
Mac - Thanks for the correction - that’s the structural engineer in me, which is more accustomed to dealing with squared-off shapes and other businesses where ‘‘wall’’ is the term of art. Another railroad/ tank business idiosyncracy . . .
When I was doing contract work on the in-plant track network at Lukens Steel in Coatesville in the 1975 - 1980 time frame, I saw a lot of the forging molds or dies for the pressure tank and tank car heads and ‘head shields’ that were then being retrofitted to a lot of the tank cars for added safety = resistance to punctures by adjoining couplers during a derailment, etc. I don’t recall seeing many of the actual heads and head shields where I was at - but the molds were typically 2’’ to 3’’ thick, if I recall correctly. If they had to be that thick to impart and contain the forging or pressing forces of the head and head shield without distorting themselves, then I surmise that the heads and head shields were in the 1’’ to 2’’ thick range. Does that sound anywheres near right ?
I was thinking exactly that as I was reading through your post - ‘‘Well, at least that crushed limestone would have been pretty good to neutralize the acid’’ - and t
Not really my foreman lost his job. He loaded a Deadlined tanker and overroade a certified inspector that the trailer was safe. AKA ME at the time I was the one and ONLY person in the shop certified to do thickness and any and all repairs to the Hydraulic componets of the tanks. Well the trailer in question the driver before this incident had reported a leaking Saftey valve and I concurred and Deadlined the trailer. Needed a replacement Valve and Flange on the trailer. Foreman told me I was wrong wrote repairs not needed. DOT was like were is your certification to make that statement. He did not have it I was the only one and therefore he was in Trouble. Cost the boss IIRC 15K in fines 40K for the repairs to the trailer afterwards namely new axles and such. Funniest thing was Bossman made the Ex-foreman PAY FOR THE FINE out of his 401K and the court agreed with him.
Let’s do a propane tank car. Internal pressure is 350 PSI or so. Diameter is 10 feet (120") or so. Allowable stress on mild steel is, say, 12,000 psi (allows some safety factor.)
The cylinder wall is a hoop and pressure stress on any point can be calculated as such.
Starting with a 1" long hoop.
120 x 350 = 42000 lbs stress on both sides, 21,000 lbs stress on 1 side.
At 12,000 psi allowable, 21,000/12,000 = 1.75 inches thick.
If the shell is made of mild steel for this application that is the absolute minimum shell thickness there should be. Mechanical loading, impact resistance, and train tension go beyond this. Figure a guess of 2" thick for mild steel. Maybe less for some alloys.
Mild steel is a pretty forgiving metal. It generally doesn’t shatter. and is very forgiving of heat stress and welding. Many alloy steels have higher tensile strength ratings but don’t do so well with impact or crack propogation. They may also be harder to weld reliably.
The ends and any tapered parts are a whole different calculation.
What it comes down to is that wall thickness depends on what the tank is designed to haul.
Redore - thanks for that part of it. As I surmised, the 1.75’’ to 2’’ thickness needed for that pressure is a whole lot more than for the simple train forces of tension and compression that I was working with - say, 1/2’’ or so.
Over lunch I thought about the analysis of the short-cylinder beam body of the car from a weight-carrying perspective. To do that right, I need to find the proper formula for a cylinder whose axial length is about 4 or 5 times its diameter - by right, I ought to be able to derive it from scratch, but that kind of thing was 35 years ago . . . So, as a greatly simplified ‘1st cut’, I considered a 200,000 lb. load concentrated in the middle of a 40 ft. long car. That results in a ‘bending moment’ of 2,000 KIP-ft., or a ‘couple’ of tensile and compression forces at the top and bottom of the 10 ft. diam. car shell of about 200 KIPs or 200,000 lbs. Going with the same 1/2’’ thickness and your 12,000 PSI figure, that means each inch of width of the car could carry 6,000 lbs. of that force, or that about 33 inches = 3.8 ft. of width would be needed. That seems roughly OK for the top and bottom segments of a circumference of about 31.5 ft. - a little over 10 % of the hoop. But of course, a much closer and more sophisticated and detailed analysis is needed to be more definitive about this.
Nevertheless, considering those 3 loading conditions, the pressure loading is clearly the most demanding or ‘governing’ one, requiring the 1.75’’ to 2’’ thickness; it looks like the train forces and the c
For the tank, it is indeed a shell. For a cylinder floating in space, it is the cylinder wall. 
Paul, don’t forget the car has to take the forces in buff, too, so you’d have to do a rather icky buckling calculation. Also, remember that the load is applied through the stub sills, not uniformly around the cylinder. Finite element analysis anyone?
Redore,
While my CFR is packed away due to a recent move, your speculations as to propane tank walls are a bit off. My recollection is head and shell thickness of Propane/Anyhdrous Ammonia 112A340W was/is about 3/4 inch. In this case 340# is the hydrostatic test pressure, done every 5 years. Saftey Valve Start to Discharge pressure is 280.5#. Design Burst strength is 850#.
Paul - When the A cars were retrofittted to S, T or J, each of which involved adding a headshield, the headshields were 1/2 inch as I recall. Last I looked the half inch is what is required today for new construction of Spec 112 cars.
Mac
[
Yep - see above. However, it’s not a long, slender column, so I believe the customary Euler 2nd degree differential equation for buckling isn’t applicable. Given the low values/ state of the superimposed stress field for that, I doubt if it would control the design, or even come close. I do remember analyzing a long thin intermodal car frame for that, and being surprised at how tough it was to make it all work out.
I also thought of the analytical challenge of the stub sill connection, too. All this reminded me of my ‘Theory of Elasticity’ class - spring semester, junior year - toughest class I ever took. Prof. Viscomi will be proud that I thought of it - which might make up for my performance on his exams. Good catches anyway, Don.
- Paul North.
From personal experience riding the end of tank cars with head shields, yes, they are ½" thick, which at first glance seems to be a little thin, till you see one that has taken a direct in line hit.
The shield does what it is designed to, it absorbs the impact, the bracing buckles, (like an automotive crumple zone) and the outer shell might get a dent, but the inner tank retains integrity.
If you click on the link to Trinity Railcar’s website in my first post, you will find the following information for LPG and AA type tankcars. Trinity Railcar is a tankcar manufacturer.
33690 Gallon, DOT 112J340W: 0.608" thick shell and heads
33560 Gallon, DOT 112J400W: 0.714" thick shell and heads
33748 Gallon, DOT 105J300W: 0.5625" thick shell and heads
34302 Gallon, DOT 112J500I: 0.89" thick shell, 1.03" thick heads