Pic from June 1966 Scientific American
How do they deal with this? If they cut the rail, won’t it spring back to straight? Or are the bends sharp enough to kink the rails? Is it possible to kink rails, or will they break before they kink?
All steel alloys have essentially the same modulus of elasticity with slight variations but will have significantly different yield, ultimate, and shear strengths as well as elongation (strain) at failure depending on the formulation, heat treatment, and surface condition. I’m having some difficulty making out the scale but it appears that the lengths are 700 feet which have been compressed into this sine-wave shape by friction with the ground and even distribution of stress gradients within the rails. There is certainly residual stress in the rails due to the ‘pre-load’ of being compressed in this fashion and cutting them through a mid-section would be ill-advised. There is likely some amount of plastic deformation that has occurred during this event.
Having dealt with large orders of plate and structural steel but never anything exceeding 40 feet in longest dimension, I can only guess about how someone would load and unload these rails, much less how to try to recover this load. It seems like a nightmare but shipyards deal with awkward and irregularly shaped steel sections all the time so I’m sure the experience is out there to figure it out.
Stranger
Interesting. If those bends are trying to straighten out it should be possible to put one flat car at a time back on the track and back it up, eventually getting the whole train and rail segments back together. The original train could navigate fairly tight curves and those rails would bend and unbend repeatedly without fatigue, but there would be a limit beyond which the rails would deform, at least some. It seem likely that the extremely small radius of those bends would cause the rail to deform by twisting, which would probably mean they’d have to be scrapped.
The “700 feet” in the caption is probably wrong – rails were (are?) typically welded into quarter-mile strings before being loaded on the train.
Such trains go around curves of 500 ft radius, anyway – dunno how much sharper they can handle.
That’s an interesting photo in link in the first post. As a railroad enthusiast who has looked at thousands of train photos I have never seen something like this. I wonder if they were able to salvage the rails or if they were too bent up to be used again.
Rails are surprisingly flexible. Here is a video of a train load of rails going around a curve: https://www.youtube.com/watch?app=desktop&v=O4kC262exiw
Photos and article about transporting very long rails here: https://www.voestalpine.com/blog/en/innovation-technology/railways/the-logistical-challenge-of-transporting-long-rails/
That was incredibly satisfying to watch! At first I was like “ok, so it’s a train going around a corner…” and then all of sudden it was like, “wait, what!!”
Yeah.
Each of the cargo rails is resisting being curved. The total force of however many rails being curved is being applied as a sort of “centrifugal-equivalent force” trying to push the railcars off the track to the outside.
Somebody had to do the math to make that work. The train of course is traveling slowly which helps.
Maybe. I don’t think the industry got from 50 foot sections to 1/4 mi ~= 1300 foot sections in one bite. Back in 1965 it may well be that 700 feet was the state of the art. That was 60 years ago now. Technical progress marches on just as time does.
The bends are trying to straighten out some. It’s a good bet the areas of tightest curvature have exceeded the elastic limit and the at-rest no-stress condition of that part of each rails is curved, not straight. OTOH, the relatively straight or lightly curved sections between the “kinks” are probably still within their elastic limits, and if released to relax ought to return to real close to straight. Net of any adverse twist as you noted.
Which suggests to me you could recover the train by pulling the first relatively straight section onto the running track as you suggest, then cutting loose the badly bent section just behind. Lather rinse repeat until you’ve eaten the whole train.
But just like felling trees, there’s insane levels of force stored in that deformation and you’d want to do any cutting using remote machines, not men holding torches.
Offhand guess: the curve in the video is maybe an 11-degree, and CWR trains probably go around main line curves that sharp with no fuss. But maybe the Deshler curve isn’t top-notch track and they needed to look it over beforehand.
That’s amazing! I wouldn’t have thought this was possible. The drag this must be putting on that single locomotive effectively bending all those many rails at once must be incredible. It’s also curious that the resistive forces don’t derail the cars.
And, related to the flexibility of steel rail, I was wondering how track is laid to conform to all the different necessary curvatures in a rail line, and was just watching some YouTube videos on how that’s done. Rail segments are made and shipped as straight “sticks” of whatever length, and are bent in place into whatever curvature is needed. In the old days and maybe to some extent still today, iron bars can be used to bend rail into shape, especially for fairly broad curves. Or power equipment can be used. For sharper curves like in rail yards, special hydraulic presses are used. According to one dude, the only track that may be pre-curved is that for city transit lines, where curves might need to be very sharp.
I still don’t understand how they were able to solve the buckling problem (due to CTE) with continuous rail. I watched a couple YT vids on the subject a few years ago, and they briefly mentioned “pre-loading.” But completely glossed over the details.
CTE?
Coefficient of thermal expansion.
Stranger
THX. 
Windmill blades loaded on flatcars are pretty cool too. Saw one recently that each blade spanned 3 cars with the mount on the first and third cars. Car #2 was just a coupler really. And they seemed like extra long flat cars.
Quite the trick.
I was curious about this, too, and this is what I got from a couple of videos, most notably the one I linked here. The basic answer, to my understanding, is that although materials expand according to their coefficient of thermal expansion, if the material is under sufficient compressive stress, there is an opposite effect called elastic deformation . Simply put, AIUI, continuous welded rail (CWR) constrains thermal expansion (within some limits) according to the elastic modulus of the rails, as demonstrated in the video with the experiment with a steel bar held in a hydraulic press and then heated with a blowtorch. It exerts a lot of pressure on the press but expands very little.
The linked video doesn’t mention it, but I believe that “pre-loading” refers to the practice of stretching CWR as necessary so that the rail is the length it would be at some standard nominally warm temperature, called the stress-free temperature (SFT) and then welding in place. This means the rails don’t go into compression until the SFT is exceeded, and minimizes the enormous compressive forces that would otherwise be created on hot days.
OTOH, that means the pre-loaded rails will be in eqaully enormous tension whenever it’s colder than that SFT. In areas with large annual temperature swings that too would be an issue.
Dead straight runs don’t really have a problem with tension other than putting lengthwise loading onto each place the rail is anchored to the ground. But in a curve, rails under tension will both want to pull sideways towards the inside of the curve.
Just one more semi-dynamic load to put into all the calcs necessary to build this stuff, but SFT is essentially a way of “splitting the difference” so the rail is under either tension or compression nearly all the time, but never too much of either.
But when they do need an expansion gap they have breather switches…. Usually found at transitions to jointed rail or bridges.
“Splitting the difference” is exactly what it is. It’s amazing how intricate railroads really are. It’s hard to believe that the first transcontinental railway was completed across Canada in 1885, and in 1869 in the US, and that it worked as well as it did with the limited knowledge and primitive technology of the time.
One interesting factoid about SFT numbers is that they’re surprisingly much higher than the average ambient temperature in the area. This isn’t necessarily because compression stress is more damaging than tension stress, though it may be a factor. It’s mostly because it’s really surprising how much direct sunlight can raise the temperature of a rail above ambient, and because the effective SFT value goes down over time as stresses build up in the rails.
There was a major derailment of a freight train in Napadogan, NB, in 2021, and the Transportation Safety Board of Canada (TSB) report about the accident provides some interesting technical insights. Napadogan has an average summer high temperature of around 24-25°C (75-77°F), and the average low in winter is about 10-11°C (50-52°F). But CN’s preferred SFT in that area is 32°C (90°F). The tracks in the area were in the process of undergoing de-stressing, a necessary process because compression stress builds up over time and the SFT (or what the TSB calls the preferred rail neutral temperature) goes down over time.
Quick summary of the findings:
-
Degraded condition of rail anchoring on the non-destressed sections reduced the strength of the track and its resistance to movement.
-
The ambient temperature at the time was about 25°C, but periods of bright sunlight likely increased rail temperature beyond 40°C (104°F).
-
The combination of compressive thermal stress and the longitudinal forces from the braking train on a track with a lower rail neutral temperature and degraded rail anchoring likely initiated the track buckle.