Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
How to Derail a Train
After I graduated with a degree in Mechanical Engineering I went to work for FreightMaster, a company that manufactured end-of-car cushioning units for railcars. The units are huge shock absorbers that are installed behind railcar couplers to help cushion the “lading” (i.e, the stuff being transported). We would put strain gauges on the shafts of these shock absorbers and smash railcars together to test new designs. Very cool.
FreightMaster also sold simulators used to train locomotive engineers. A simulator consisted of a minicomputer attached to a control panel from a diesel-electric locomotive. A monitor connected to the computer told you how fast you were going, what your brake line air pressure was, whether you’d just gone off the rails, and so on. As part of the deal, FreigtMaster would “digitize” the routes that the engineers would someday drive and feed the routes into the simulator.
I did a bit of programming for the system and got to play a locomotive engineer. In the process, I found out how easy it is to derail a train. The most obvious way to do that is to take a curve too fast and let inertia do the rest. But there are more subtle ways to make it happen.
For example, if you accelerate on a curve, you put the train – which is like a long, stiff string – in tension. Too much tension, and you pull the string off the inside of the curve. Conversely, if you brake while in a curve, you put the string in compression. Too much compression and you push the string off the outside of the curve.
The strangest way to derail a train is via “resonance.” Resonance occurs when all the oscillations in a system reinforce – rather than offset – each other so that the system oscillates at a higher amplitude. If you’ve ever seen a video of the Tacoma Narrows Bridge collapsing, you have an idea of what resonance can do.
Trains tend to “rock” on American tracks because of the way in which the rails are laid. Rails are 39’ long so that they will fit inside a 40’ gondola car. When rails are laid, the joints are offset by half a rail, so that there is a joint on the track every 19.5’ – first on one rail and then on the next. The weight of the trains traveling over the joints causes them to “work,” and they end up being a bit lower than the rest of the track. As the cars roll over the joints, then, they dip a bit. This means that the cars “rock” as they dip first to one side and then to the other.
At 17.5 mph, the cars hit resonance. If they stay in resonance long enough, they can actually hop off the tracks. Therefore, the engineer must pass through 17.5 mph – whether he’s accelerating or decelerating – as quickly as possible.
Published in General
You wouldn’t happen to know a Gomez Addams, would you?
But not so quickly on a curve that you exceed tension or compression limits.
That one is surprising. It make perfect sense, but I’ve never considered it before.
Tacoma Narrows Bridge, shortly before the collapse:
The fun part is when you realize that when you do want a train to derail, a lot of time it won’t cooperate.
Like the Crazy Eights runaway train (the real event that inspired the Denzel Washington movie Unstoppable). In real life, they actually did try to stop it with a portable derailer, but the train blew right through.
At the point they tried to derail it, the train was going very fast. Trains are massive and the inertia must have been incredible. So, it’s not that surprising that the train kept going.
While I was with FreightMaster, some of our folks were working with a railroad company in the Rocky Mountains. There was a terrible accident while they were there – fortunately not the train that was pulling our instrument car. The engineer was inexperienced and he’d bled off too much of the air in his brake line by the time he’d gotten to the top of the grade. All he had left on the downgrade was his dynamic brake.*
When the grid melted, his train became a runaway. They radioed the train that was ahead of him to warn them. The leading train sped up, but it was too late. When the runaway hit them, they were doing over 90 mph. The engineer in the runaway was killed as were the people in the leading train’s caboose.
One of our guys saw the locomotive from the runaway after the accident and said that it looked like a crushed beer can.
*A diesel electric locomotive has a huge diesel engine that powers a generator. The generator, in turn, powers electric motors at each of the driving wheels. For dynamic breaking, the motors are switched so that they become generators. The force needed to turn the generators helps slow the train. The electricity generated is fed to a big grid at the top of the locomotive. Huge fans help cool the grid.
Trains are just plain weird.
For example, for long trains, you have to put it in reverse for a bit before going forward. That way, you compress all of the couplers, so when you start out, you have a little bit of slack to let the engine get some inertia going. That adds mass a little at a time until the slack comes out of all of the couplers, instead of trying to make the whole train start at the same time.
Slopes are another thing – you don’t normally think of a four percent grade as being “really steep,” but right now, the steepest incline in the US railway system is right at 3.3% – one foot of rise in 30 feet of track. There’s not a lot of tracks with more than a two percent grade. That’s why you have things like the Tehachapi Loop, where a train can often be seen looping over itself to avoid going over a two percent grade.
https://en.wikipedia.org/wiki/Tehachapi_Loop
Another interesting thing about train engineering; a classic from Richard Feynman:
I think it’s a guy thing.
My great-grandfather was a wrecking engineer.
While working in the Rockies, Ken, one of our people was riding up front with the locomotive engineer. Ken said that his heart was in his mouth the whole time because people kept trying to beat the train at railroad crossings. He said that a couple of times they missed the cars by only a few feet.
The engineer said he’d hit a car just a few days before. The driver – it happened to be a woman – almost made it, but he knocked her rear bumper off and spun the car around. He slammed on the brakes, got the train stopped (after about a half mile), and went back to make sure that she was okay.
When he walked up to her, she shouted, “Why the Hell didn’t you swerve?”
I don’t have a lot of experience with passenger trains in the US. I rode quite a few while traveling in China. My favorite experience was on the MAGLEV train from the Pudong Airport to the city center (Shanghai). I was like a kid when I got off, “let’s do that again!”
Almost got a Darwin Award and fully deserved one.
We have a commuter train here in Orlando, and on a regular basis, people either try to beat the train (and fail miserably), stop on the tracks (with predictable results), or get hit while walking down the tracks.
Now, these are reasonably quick trains, doing 70 MPH or so on long stretches, but they’re loud enough to hear the engine noise alone from a quarter-mile off, and they lean on their horns at the slightest provocation. It’s not like they’re a surprise, is what I’m saying. One time, I checked, and it takes a grand total of about six seconds to get from “clear of the train” on one side of the tracks to “clear of the train” on the other at a normal walk.
I’ve also seen a guy in a very nice Lamborghini decide to beat a slow-moving train by driving under the crossing arms, which were all the way down. He made it. Mostly. Apparently, his height judgement was not perfect, since the arm scraped the whole length of the roof of the car. Since it was a Lambo, it was probably a few thousand dollars to get the paint redone.
“Trains tend to “rock” on American tracks because of the way in which the rails are laid. Rails are 39’ long so that they will fit inside a 40’ gondola car. When rails are laid, the joints are offset by half a rail, so that there is a joint on the track every 19.5’ – first on one rail and then on the next. The weight of the trains traveling over the joints causes them to “work,” and they end up being a bit lower than the rest of the track. As the cars roll over the joints, then, they dip a bit. This means that the cars “rock” as they dip first to one side and then to the other.”
This is no longer the case. Jointed rail has mostly been eliminated on US mainlines for quite some time. Instead, what is sometimes referred to as ribbon rail is used. This is lengths of rail welded together to form continuous rails that can be as much as a half-mile long. Resonance is indeed an important consideration, but rail cars are not particularly even multiples of 39 feet, and it is the spacing between the wheels that would matter.
True. I was working for FreightMaster back in the late 70s and early 80s and welded rail was starting to replace bolted rail. With bolted rail, you can leave a gap between the ends of the rails to allow for expansion. With welded rail, you’ve got to anchor the rails down so that they won’t expand, otherwise the rails will buckle or even pop out of the rail bed when it gets hot. That’s another great way to derail a train:
Since it was a Lambo, they could probably afford it.
I lived near a train yard for awhile, and wondered what all that business of going backwards was. Thanks for the explanation–I kind of sort of get it in a vague way. :-)
A skillful engineer will stop the train so that the locomotive stops first and the cars keep inching forward until each runs into either the locomotive or the car in front. That way, when the train starts again, the locomotive has time to gain a bit of speed before the couplings all run out and it’s pulling the entire weight of the train.
Before I joined FreightMaster, I thought being a railroad engineer was the easiest job in the world. Just sit in the cab and wave to the kids as you go by. Everything is simple until you know something about it.
I can relate. All the casual “learn to code!” talk makes me go cross-eyed.
I get it! It clicked.
Yes, cool!
The Rochester (NY) Institute of Technology has a program in packaging science. Their lab was always fun to visit at the college’s annual open house. They had setups to simulate rail cars banging into each other during coupling operations, tractor trailer rides across the country, and handling by forklifts, humans, and other transporters. Railcar coupling and truck transport are surprisingly violent activities! The RIT lab tests packaging design to see if the television or medical diagnostic equipment, beverage containers, and other stuff being transported would arrive undamaged. Fun discovery – a small package (Amazon, USPS, UPS, FedEx) is dropped three times from a height of about 3 feet (waist high) during its transit from the warehouse to a house.
I’ve been on at least 4 trains a day, 5 days a week, for over 30 years and did not know this! Thanks!
I assume vertical flexing of the rails as the weight of the train passes over must still be desired, since I still see new timber (wood) ties used (and regularly replaced) rather than presumably more durable concrete ties on main lines across west Texas and New Mexico.
I don’t think that wood ties can be used with welded rail (though I could be wrong). I don’t think that they’re heavy enough to keep the rails from expanding and contracting as the temperature changes. The welded rail I’ve seen uses concrete ties.
My guess is that wooden ties and bolted rail are still used because they’re less expensive to buy and install (probably more expensive to maintain, though). And they’re fine for relatively low-speed rail.
A warning: discussions like this are what triggers people into watching train videos on YouTube.
For hours.
And hours.
The vibration and centripetal force of trains going around curves would gradually pull the tracks and ties out of alignment. Work crews called “gandydancers“ would periodically have to re-position the tracks and ties. The work is now done by machines. But back in the day it was muscle and sweat and rhythm. The chants and songs and “hollers” of the foremen kept the crews in time and working together – like the coxswain in a boat. They heavily influenced The Blues.
https://youtu.be/c1O2X890tig
I don’t think it’s possible to stop rails from expanding and contracting. I think you have your engineering wrong.
There is something going on there, but I don’t remember how it works. I remember when I lived in Phoenix and they started installing light rail several years ago. There were actually some semi-technical articles in the newspaper about how the rails were connected with special welds and stuff. And how things could come apart from expansion/contraction if it wasn’t done right.
And I remember in High School geometry class, one example of “unexpected things” was the problem of a half-mile-long-or-something piece of railroad track, that was 1″ too long. If it was ‘bent’ to fit in, how high would the gap be in the middle? Turns out it was several feet, as I recall, more than enough to walk under while standing upright.
Sort of like a country, not welded together right.
You can keep the rails from growing longer (or shorter), but you can’t keep them from expanding in every direction, so they’ll get “fatter” (or “skinnier”). Here’s the deal. Let’s say that you can stretch a rail by one inch by exerting a longitudinal force on it of, say, 2000 pounds (making up the numbers, of course). Now let’s say that we’re building a railroad in an area where the temperature can range from a low of -20 degrees to 120 degrees – so a swing of 140 degrees. If raising the temperature of a rail by 140 degrees will increase its length by two inches, then we can keep it from lengthening by exerting 4000 pounds of compressive force on it (assuming a linear coefficient of expansion). Naturally, we’ll throw in a big safety factor.
Pipelines have the same problem. We can either lay a surface pipeline in a zigzag fashion to allow it to expand, or tie it down securely enough that we can offset the pipe’s tendency to expand as the temperature or internal pressure rises. If we’re burying the pipe, we’ve got to clamp it down, otherwise it will force its way out of the ground at an elbow. We can do this with thrust blocks (chunks of reinforced concrete).