Tuesday, 10 June 2014

Ever wondered why Continuous Welded Rail doesn't need expansion gaps?

Prior to the widespread use of Continuous Welded Rail (CWR) on UK railways, track was laid in 'panels' rather like pieces of model railway track, but in sixty foot lengths. These panels were each bolted to the next using a pair of 'fishplates' as shown below. One plate is placed either side of the rail and four bolts hold the plates and rail in place, with an expansion gap left between the rails. As the rail expands in warm weather this gap closes, and the gap is wide enough that on all but the very hottest of days the expansion can be accommodated. If the rail expands beyond this point, the track will buckle.

A fishplate joining two track panels. Note the expansion gap left between the rails.

Traditional sixty foot track lengths produce the familiar (less so these days) clickety-clack sound of the train wheels over the track joints. Unfortunately, fishplated rail ends are less 'stiff' than the rail itself, so the rail tends to 'pump' up and down at the joints under the weight of a passing train. As it pumps, it allows rainwater to penetrate beneath the sleepers at the rail ends eroding the trackbed and making the pumping worse. Also, the rail ends wear as the wheels pass over them, and the bolts need regular checking and re-tightening. All of this means that traditional track joints require a lot of maintenance. To obviate this, and to ensure a more accurate and reliable track alignment, jointed track has now been almost entirely replaced by Continuous Welded Rail (CWR) on UK railways.

Old and new; A 1930s A4 Pacific on modern Continuous Welded Rail

CWR therefore requires less maintenance than jointed track, and will remain in accurate alignment for longer. But if it is continuous, with no expansion joints, how does it cope with thermal expansion in hot weather?

Its laid it's in tension by hydraulically stretching it before cutting to length and welding to the previous length. Sufficient tension is applied to achieve a 'Stress Free Temperature' (SFT) of 27 degrees C (i.e to relieve the tension by expansion the rail would need to be heated by the sun to 27C, whereupon it would neither be in tension or compression).

The rail temperature at which the expansion stress has the potential to start causing buckles on normal, plain track in good condition is 32C above stress free temperature. So, an SFT of 27C means that rail temperature has to reach 59C before mitigation measures are made (such as blanket speed restrictions). And a rail temp of 59C in high summer, with blue sky, (worst case) requires a shade air temperature of 41C, something which has never been recorded in this country.

In summer the effect of the sun lifts the temperature of the rail by between 8C (full cloud cover) and 18C (blue sky) above ambient shade temperature. 

So CWR provides a quiet, smooth ride in comparison to jointed track, especially for trains that run at high speed. It also retains its alignment for longer, is most unlikely to buckle even in the hottest UK weather, and requires less maintenance. No wonder that jointed track has all but disappeared except on heritage railways where maintenance is by volunteers, and the passengers enjoy that nostalgic clickety-clack of the carriage wheels over the rail joints. 



  1. Indeed; now I know, also - thank you, Vince Chadwick

  2. Well now I know too. I had always thought that the longitudinal joints were there to take up expansion but now realise that these are more likely for track circuiting than expansion.

  3. Good bit of info. Two things to add:
    1) Rail (stressed or not) is also subject to low temperature stress as well. Each installation of CWR is calculated and customised to the region and climate and time of year in which it is installed. It can be a pretty complicated process.
    2) CWR also has stabilisers attached which bind the rail from sliding along its length due to thermal forces. These are called anchors and are what are primarily responsible for keeping the stress in the rail uniform and within tolerance.

  4. Thanks for that extra info, Jovet.

  5. Excellent, well done. Very informative.