Many many thanks Bald Rick. Very informative.
There was an article in our newspaper today about this. And it discussed tensioning the track. And that sometimes they cut pieces out of the track and reweld??!!
I just discovered that the greatest temperature variation in one place during a 24 hour period was in Browning USA and was 56 degrees Centigrade. Goodness knows how rail track could cope with such an event.
Almost all Continuosly Welded Rsil (CWR) is ‘tensioned’ or more specifically stressed. In this country, when it is installed, it is pulled (stressed) by hydraulic jacks to a length that it would be naturally if it was at a temperature of 27C (known as the Stress Free Temperature, SFT). This is a simple calculation, based on the rail temperature at installation, the length of rail being installed, and the coefficient of expansion of steel. While the jacks and clamps hold the stressed rail in place, it is then clipped down, and the ends welded to existing rail. The weld cools and gets sufficient strength very quickly.
Thereafter, if the rail temperature is below 27C, it is effectively under tension without moving as the steel tries to contract but can’t; the colder it gets the more stress is in the rail. When rail temperatures get below about -5C the increased stress leads to a higher chance of broken rails, as rail defects are effectively weak points in the rail and the stress can overcome them. This is why broken rails are more prevalent in cold spells.
If the rail temperature is over 27C, then the rail is effectively in compresssion, as it is trying to expand but can’t. The rest of the track system (clips, sleepers, ballast) hold it in the correct position. When it gets very hot, the steel wants to expand so much that the forces will eventually find a weak spot, usually on a curve or geometry defect, and the rail will expand to the outside of the curve, ie a buckle. For good track that has not been worked on, the risk of this happening increases when the rail temperature is 32C higher than the SFT or more. That is 59C for properly stressed track, and that means a shade air temperature of 41C in high summer, which we have never had in this country.
The issue with heat speed restrictions occurs when the SFT is lower than 27C. This can only be because it was installed at a lower rail temperature and has not been fully stressed, or the stress is not known. Added this, can be some other characteristic of the track system that means the track does not have the usual level of restraint (typically ballast that is not consolidated, but can also be missing ballast, wide sleeper spacing, tight curves, type of track components, and many other factors). Occasionally records are missing in which case you have to assume the worst case for the given site until the stress can be tested. A simple calculation of all the applicable factors, along with the SFT, gives you a Critical Rail Tempersture (CRT), which when reached means some action must be taken:
1) a watchman is deployed to maintain constant vigilance on the section of track concerned, who can report to the signaller immediately there is any sign of track movement. This is known as the CRT (W).
2) a little warmer, and a 30/60 (30 for freight, 60 for passenger) speed restriction is imposed. This is CRT 30/60.
3) warmer still and a 20mph restriction must be imposed. This is CRT20, and the most restrictive action short of blocking the line.
The speed restrictions do two things, a) reduce the forces on the rail, particularly the force on the inside of a curve as a train takes it, and b) reduces the consequences of a derailment if one occurs.
The worst case for CRTs is installing new track on a frosty April night, followed by a warm sunny day. This can easily lead to an ambient temperature swing of getting on for 30C, with a rail temperature swing of 40C or more. That is a guaranteed CRT20 speed restriction, sometimes even if the rail is stressed.
Reputationally even worse is installing new track on a very cold late Feb, early March night, followed by a sunny day. If the rail isn’t stressed on the night of installing (which often happens), then you can get a ‘heat’ speed restriction with the ambient temperature barely above freezing. That takes some explaining to the Ops team.
Finally, re cutting sections of rail out. That’s not usually done here except in a couple of unusual circumstances.
Firstly, if the weather warms up quickly during rail installation then it is quite possible that the rail will ‘grow’ by several hundred millimetre post installation, and it will need cutting again to get it to fit before stressing. (It is not usual practice to install new rail above 27C or so unless it is an emergency, but the weather can catch you out. I had one job where we had to cut it two or three times; it then cooled quickly and it shrank too much!)
Secondly there are a handful of locations around the country that need different rails for ‘summer’ and ‘winter’ - swing bridges. The size of the swing bridge is (hopefully!) fixed, but the rails will expand / contract. Therefore to meet the tolerances necessary to get everything new to fit and swing, slightly shorter rails are installed from around Easter to October. But the temperature makes them roughly the same length as the winter rails in colder temps, if you see what I mean.