Why Railroads Don't Need Expansion Joints
[Note that this article is a transcript of the video embedded above.]
One of the most common attributes folks imagine when they think of trains is the clickety-clack sound they make as they roll down the tracks. The thing is, most trains don’t make that sound anymore. Or really, I should say, most rails don’t make that sound anymore. Trains are still pretty clickety-clacky, but they’re far less so than they used to be. And here’s why: those rhythmic clicks and clacks came from joints in the tracks. Those joints were a solution to a transportation problem: you can only roll out a length of rail so long before it gets difficult to move them around. It’s easier to have short segments of rail that can be bolted together in place. But, they were also a solution to a thermal problem.
You might be familiar with the idea of an expansion joint: a gap in a sidewalk or handrail or bridge deck or building meant to give a structure room to expand or contract from changes in temperature. I actually made a video on that topic a few years back. The joints on railroads were bridged by fish plates, but with a gap, so on hot days, the rail would have room to grow. But look for a joint on modern railway, and you might have a hard time.
We’re in the middle of a deep dive series on railway engineering, so don’t forget to check out the other videos after this one. A lot of new track these days uses continuous welded rail or CWR that eliminates most joints. Large structures subjected to swings in temperature that don’t account for thermal expansion and contraction can run into serious problems or even fail. So how do modern railways get away with it? I have a bunch of demonstrations to show you. I’m Grady, and this is Practical Engineering. Today we’re talking about continuous welded rail.
As much as I enjoy a good conspiracy, the railroad companies don’t have access to some kind of special steel that doesn’t expand or contract. Rails really do experience thermal contraction and expansion. In the US, they would be installed in roughly 39-foot sections. In general, tracks would be laid out so that on the hottest days, the gap between sections would just barely close. But, this style of jointed rail (although it solved some of the practical problems of railroad construction) had some serious drawbacks, too. First, it was noisy! The famous clickety-clack of railroads was caused by each wheel passing over each joint on the track. I’m a simple man. I grew up listening to that clickety clack, or as they say in Korean, “chikchikpokpok”. It brings a certain nostalgia. But when you consider how long a train is, and the fact that most cars have at least eight wheels, and that train journeys can be hundreds of miles long, that’s a lot of clicks and clacks.
The railroad companies might say too many, because noise is just a symptom. Each time a wheel clacks over a joint in the rail, that impact batters the steel, eventually wearing it down at each location. Try as they might, railroads could never make these joints quite as rigid as the rest of the rail, meaning that (in addition to the extra wear) they would create additional load on the ballast below, and the flexing would cause freight cars to rock side-to-side in a phenomenon called rock and roll. All this creates a maintenance headache, increasing the cost of keeping railroads in service. And it’s why most modern railroads use continuous welded rail: it’s a huge reduction in the maintenance costs associated with the wear and tear from joints. In CWR, rail segments are welded together using electric flash butt welding, arc welding, or in some cases, THERMITE welding. These welds have much higher stiffness than the old joints and, of course, are ground smooth, so they lack clickety clacks. But they still expand and contract with changes in temperature like most materials do. Let me show you how this works.
I’ve set up an aluminum rod on the workbench with one end clamped down and the other free to move. I put a dial indicator at the end so we can observe even tiny changes in the length of the rod. You can see on the thermal camera that we’re already starting at a fairly warm temperature; that’s Texas for you. But, rather than wait for the weather to get even warmer, I’ll speed things up with my sunny day simulator. Notice the dial on the indicator climbing steadily as the heat is applied.
This is an example of unrestricted thermal expansion. That just means nothing is keeping the rod from growing under the increase in temperature. And, engineers can predict the change in length from most materials with a pretty simple formula. Multiply the difference in starting and ending temperatures by a coefficient of thermal expansion that’s easy to look up in a table. This aluminum rod expands by about 0.002% for every degree celsius it increases in temperature. Steel is about half that. Structures like bridges with expansion joints and jointed rail are designed to allow unrestricted thermal expansion. When the hot day comes, the materials expand into the gap. That’s usually a good thing. The structure doesn’t build up stress and stress is what breaks things. But, part of the reason CWR can get away from expansion joints is that changes in temperature aren’t the only way to change the dimensions of a material.
I’ve set up another demo using that same aluminum rod. This time I put it inside this length of pipe and put a nut and washer on both sides. I put the dial indicator on the end, just like before. Now, watch what happens when I turn one of the nuts. Well, if you’re not careful, the whole rod twists. But if you can keep the rod centered in the pipe, and the nut on the other end from twisting, you can see the dial indicator registering the rod getting longer. There’s no change in temperature here; this is a totally different phenomenon: elastic deformation. Turning this nut applies a tension force to the rod, and it stretches out in response.
Just as all materials have a mostly linear relationship between temperature change and length change, all materials also have a similar relationship between stress and change in length (often called strain). If you stress a metal too far, it will undergo a permanent (or plastic) deformation. But within a certain range, the behavior is elastic. It will return to its original length if the stress is removed. And just like the slope of the line for thermal expansion is the thermal coefficient, the slope of the elastic part of a stress/strain curve is called the elastic modulus. And this is part of the secret to continuous welded rail: restrained thermal expansion. You can overcome one with the other. Let me show you a demonstration.
Here you can see me using a hydraulic press in a way that’s not exactly how it was designed. First, I get this iron pipe set up in the press with enough pressure to hold it tight between the cylinder and table, about 3 tons. Then I heat up the pipe with the sunny day simulator. What do you think will happen? Will the hydraulic press break as the steel expands, or something else? Well, it wasn’t quite as dramatic as I was hoping, but that little movement in the gauge still corresponds to about a quarter of a ton of additional force in the hydraulic cylinder. You can kind of think of this in two separate steps: the steel expanded from the heat, but then the additional force from the hydraulic press unexpanded it back to its original size. The thermal and elastic deformations canceled each other out and the pipe stayed the same size. In reality, the force required to counteract thermal expansion should have been more than that, so I think the frame of my hydraulic press wasn’t quite stiff enough to hold the ends perfectly rigid. But you still get the point: you can trade temperature changes for stress and keep the material from changing in size. With a little recreational math, we can combine the two equations to get a single one that gives you the stress in a restricted material from a change in temperature.
So that’s just what railroads with CWR do: they connect the rail at each tie to hold it tight and restrict its movement, allowing it to build up tensile or compressive stress as its temperature changes. Of course, too much stress can fail a material, but steel can handle quite a bit before it gets close to that. Railway here in Texas can range in temperature from below freezing to over 100 degrees F or 40 C. That means every mile of steel wants to be more than 2 feet longer on the hot days than the cold ones. In metric, every kilometer of rail would expand by roughly half a meter, if it wasn’t restrained. Using the formula we developed here, we can see that fully restraining the rail across that temperature range results in a stress of about 15,000 psi or 100 megapascals, way below the tensile or compressive strength of any modern steel, especially the fancy alloys they use these days. But it’s not quite that simple, particularly for compression. Just because a material has a high compressive strength (and steel does), that doesn’t mean it won’t fail under compressive loading. Let me show you another demo.
We’re back to the aluminum rod, but this time I clamped both ends to create a restricted condition. Now watch what happens when I apply the blowtorch. Our equation says the rod should build up stress so that the elastic strain is equal to the thermal expansion. But that’s not what happens. Instead, the rod just deflects sideways, an effect known as buckling. Even though aluminum is relatively strong under compression, the long skinny shape of the rod (just like the rails on tracks) is particularly prone to buckling. Obviously, if a rail buckles on a hot day, it’s a pretty serious problem. The material itself doesn’t fail, but the track does fail at being a railway since trains need rails to be precisely spaced without crazy curves. Many train derailments have happened because a continuous welded rail got too hot and buckled, an effect colloquially known as sun kink. So railroad owners have to be really careful about compressive stress in a rail, and in the US, safety regulations require them to follow detailed procedures for installing, adjusting, inspecting, and maintaining continuous welded rail. One of the tricks they use to manage buckling is adding restraint. I’ve got one more formula and one more demo for you. The formula for the critical force required to buckle a structural member like this is pretty simple. Notice that the force goes up in inverse relation to the length of the structural member squared. This is much clearer in a demonstration. I have a length of welding wire, and I can apply a force with my finger that is measured by the scale. You can see it takes about 375 grams to buckle the rod. But watch what happens when I restrain the rod at the centerpoint, effectively halving its length. I can still buckle it, but it takes a lot more force from my finger. It happens right around 1500 grams, exactly what is predicted by the formula. Halve the length, quadruple the critical force for buckling. The spacing of railroad ties is really important because it affects whether or not a rail will buckle under thermal stress. And one of the most important jobs of all that crushed rock, called ballast, is to hold the ties in place and keep them from sliding horizontally and allowing the rail to buckle.
The other way railroads use to manage buckling is, I think, the most clever: just keeping rails from undergoing compression at all. Any continuous welded rail has a neutral temperature which is essentially the temperature it was the day it was installed. It’s the temperature at which the rail experiences no stress at all. If it’s colder than the neutral temperature, the rail experiences tensile stress, and if it’s hotter than the neutral temperature, the rail experiences compressive stress. The secret is that railroads use a really high neutral temperature to ensure the rail almost never undergoes compression. The Central Florida Rail Corridor has a neutral temperature of 105 F or just over 40 C. They only install rail on hot days, and if they can’t do that, they use heaters to bring the temperature up. And if they can’t do that, they use massive hydraulic jacks to induce enormous tensile forces in the rails before they’re welded together. On cold days when stresses are highest, they have to go out and inspect the rails to make sure they haven’t pulled apart, but a small break in a rail is nothing compared to a buckled track when it comes to the risk of derailment, so it just makes sense to use as high a neutral temperature as you can get away with.
Of course, you always get to the end of a continuously welded section at a bridge or older length of jointed rail. To keep the CWR from buckling at these locations, you need something more than a small gap. Instead, expansion joints on rails (sometimes called breathers) use diagonal tapers. This oblique joint allows train wheels to transition smoothly from one section of rail to another while still leaving enough room for thermal movement. And joints are also needed to break up the electrical circuits used for grade crossings and signals. So railroads often use stiff plates surrounded by insulation material to electrically isolate two sections of rail while keeping it stable in the field. We’ll cover track circuits in a future video of this series on railway engineering.
Even with its challenges, continuous welded rail extends the life of rails and wheels and makes for a much smoother and quieter ride. Even if you’re nostalgic for the soothing clickety-clack of jointed rail, it’s comforting to know that railways are continuously innovating with continuous welded rail.