Continuous Welded Rail

Older readers (and those visiting heritage train sites such as the WaterCress line) will know that a train journey was (or is) punctuated with a rhythmic sound as the train's wheels went (goes) over rail joints. Clicka-click-click, Clicka-click-click, Clicka-click-click; it was with you all the time. Now, if you notice it at all, there's a continuous swishing sound - at least on main line track. Old time railway track was made in about 60ft lengths which were joined by bolted plates, called fishplates, leaving a space between for expansion.

Now rail track is welded in one continuous length in a factory. The 60ft lengths of track are welded end to end to make a length of maybe a quarter of a mile. These lengths are transported on a train of special trucks. Believe it or not, the train has no problem negotiating the easy curves of a railway system.

"Laying continuously welded track"

When the long welded* rail arrives on site it is placed in position on the sleepers either inside or outside the existing track. At a later time the existing track is removed and replaced by the new long welded track.


*I use the term 'long welded' to denote rails of around a quarter of a mile long that are fabricated at the manufacturing workshop, and reserve 'continuous welded' for the final job.

Initially the long welded rails are joined by drilling holes in the web of the rail and bolting them together with fishplates. The holes affect the rail strength only minimally as they're on the centre of area. Later the rails will be welded to form continuous track from buffer to buffer many miles long. This raises the question of why it doesn't buckle during a change in temperature. After all, there used to be an expansion joint every 60ft and all construction work has to take temperature change into consideration.

"The de-stressing process"
The answer is that the rail is not allowed to expand or contract. It is laid at a slightly lower than mean temperature, usually at night, and carefully expanded using special heaters until marks on the rail line up accurately with marks at the trackside. They are then securely tightened down on to heavy pre-stressed concrete sleepers which in their turn are held by tons per yard (or tonnes per metre) of ballast. Although there can be considerable internal stresses present within the rail it doesn't affect the geometry of the track.

Column 1


The picture shows a Bullfinch propane gas heater of the type that was used to heat switches during freezing weather. Similar heaters were mounted on a trolley and used to heat rail when laying continuous track to marks as described above.

To prepare for destressing, the track is released from its fastenings, raised slightly and supported on small diameter rollers at long intervals. Pegs are driven into the ballast at regular distances close up to the rail head. Each peg has a scribed line on its top running perpendicular to the rail.

From the difference between the ambient temperature and the mean temperature a distance is calculated or read from a table to assess how much the rail must be lengthened at each peg and the rail top carefully scribed to show this. When the two lines coincide the rail will be at its mean temperature length.

The trolley with burners is run along the rail manually as evenly as possible and a man stands by the peg watching the marks as the distance between them decreases. When they coincide the rails are rapidly fastened down and the procedure is repeated with the next peg.

As work progresses the long rail is welded at the joints so as to become continuous welded rail. This is usually done using the Thermit method.

When track maintenance is done there are strict regulations about how much manipulation is allowable at the ambient temperature. At very high Summer temperatures track maintenance which involves lifting the track may be banned altogether. Failure to observe these precautions can unleash tremendous forces which can do considerable damage to the track and may be difficult to remedy.

Fabricating long welded rail

The notes below refer to information that I gathered whilst I was working on track maintenance equipment 45 years ago, but I believe not a lot has changed. If you know better please let me know. I'm always interested.

Long welded rail is formed by flash-butt welding standard lengths of rail together and processing the welds to achieve a continuous smooth running surface on the head.

When rail sections are rolled 30cms or so at the end is slightly bent as it comes off the roller and has to be cut off. The best practice, when ordering, is to specify that the rail ends must be in alignment with the rest of the rail, thus putting the onus on the supplier.

"Flash Butt Welding"

The most important piece of equipment is the Flash Butt Welder, as if the weld isn't good the whole job fails.

The welding process is in three stages; pre-heating, flashing and forging. The two rail ends are held in line by two powerful clamps, one fixed and the other on a guided mobile carriage.

In the pre-heating stage the rail ends are alternately brought together then moved apart, making and breaking electrical contact so that the rail ends are brought to red heat.

In the flashing stage the cycle is speeded up and the resultant flashing brings the rail ends to fusion temperature. Gases given off prevent oxidisation and the formation of impurities.

In the forging stage the two rail ends are forced together suddenly with high pressure so that they become welded together. Semi-liquid metal is forced out all around the periphery forming 'flash' that will be removed almost immediately by the Weld Trimmer

"The Weld Trimmer"

When the welded rail is released by the welding machine it is moved on by the traction rollers into the weld trimmer. It is still white hot. The rail is clamped by the weld trimmer and a set of tools that follow the shape of the rail section trim off the flash , leaving about 3mm and less on the head.

Moving the rail from the welder and trimming the rail takes only about 10 seconds.

"Further Operations"

Once the rail has passed through the trimmer it passes on to the next station which is exactly one rail length away. This is usually an automatic rail straightening machine, and on its way it passes through a cooling spray.

A further rail length on is the next station, probably to grind the head of the rail. There may be other work stations, but they'll all be a rail length apart so that each time the rail stops it will be worked on simultaneously at each work station.

Thus the throughput per weld is the time it takes to do and trim the weld plus the time taken to travel one rail length.

Around 1958, when I was in the Paris long rail manufacturing workshop of the SNCF, each weld took one and a half minutes. British Rail was taking 20 minutes a weld!

As each long rail was completed it is stored on a trackside quay until they can be loaded on to the special wagons that take them to site.

"Testing long welded rail"

When the welding machine was first started up it took some time to warm up so it was usual for a couple of test welds to be done using short rail ends. These would be tested as soon as possible, but production welding would procede provided the data obtained from the test welds were OK. Data was recorded on charts for each weld; in the event of a weld failure in the field this could then be referred to.

"Weld Trimmer"

Around 1959 Matisa UK got an order to supply French Railway (SNCF) type weld trimmers to British Railways long welded rail workshops at Darlington, Middleton nr. Oldham, and Southampton.

Although we had a set of drawings from SNCF there was still a lot of work to do to Anglicise the design before we could start manufacture. I enlisted the aid of Rexroth, the well-known hydraulics firm, and between us we designed a hydraulic system. It was more sophisticated and efficient than the SNCF version and worked well from the start.

Essentially the machine forced a pair of cutting tools in the shape of the rail cross-section across the glowing weld, cutting it to within about 3mm, closer at the head. During this time the rail was securely clamped between hard copper pads to protect the surface.

As the weld came out of the welder, the trimmer tools closed around the rail and when the weld contacted the tools the heavy frame of the trimmer would be forced back against powerful springs. The rail would be clamped and the trimming action begun. It was important to start the trim with the tools in contact with the weld so they only needed to move a short distance. As soon as the cut finished, the tools and clamp were released and the weld travelled to the next work station. Time for trimming was 10 seconds only.

We did quite a lot of research on the make-up of the tools. They were machined from good quality mild steel and the cutting edges were deposited as weld. We found the original specification was expensive, hard to deposit, tended to crack and was too difficult to work as it seemed diamond hard.

In the end I used a relatively cheap and easily available manganese rod. It was easy to work with a sharp second cut file and work hardened. The working edges of the tools had a 2mm radius so they didn't cut the weld but 'rolled' it off. This action helped to consolidate and strengthen the weld.

Unfortunately I seem to have lost my photographs of these machines; I don't suppose they're still in operation.

After Matisa UK closed down, I helped a small engineering firm make a fourth machine that I installed near Cambridge.

When I was working for Matisa, we could provide a complete long rail welding installation including buildings if necessary. They could also provide skilled personnel to manage and train operators for up to a year if needed.

At one time Matisa did this in Mauretania. I was detailed to go but couldn't get my fever shots in time so a friend went in my place. He reported that conditions were extreme with fine sand coated with salt, as they were near the sea, causing corrosion and wear to any machinery working outside.

The buildings that housed the welding plant was sealed as far as possible, and the air pressure inside the building was slightly raised to stop the sand blowing in through small apertures.