Subscribe via RSS

Welding Rail In South Yarra

There was a total shutdown of all lines past my apartment on the weekend of the 16th-17th November as Metro we're upgrading the signalling. It's well-needed to allow better head-ways for more services, and also to allow brand new 'high-capacity' trains to run. To do all this, they've had to cut in new isolated track joins. I assumed they'd just cut through the rails in-stu and clamp an insulated joiner over the top... but I assume the tension in the rails prevents them from doing this? Instead they've been busy removing a 20 metre length of track and welded in a new length that includes a bypassed insulated joiner!


It's bypassed as they don't want to break the track-circuit just yet. The rest of the infrastructure isn't in place yet, so a break in the circuit here would actually cause a 'blind' area on one of the sides. Hence the jumper cables. There's also a longer length of cables running down the current work area, to keep the current circuit in operation whilst the work is carried out. I don't really see the reason though, as there's a lot of protection at either end of the entire work area to prevent vehicles from entering.

The weather was reasonable, and I was half brain-dead from a cold, so I spent a good bit of time on the bridge near Cromwell Road, watching the professionals carry out their business. You could tell they'd done the job 100s of times before; their efficiency and precision was great to watch.

Step 1 - Align The Rails

The first step would probably be the most important in the whole process; misaligned rails would cause untold problems in the future and therefore a lot of time was spent getting the elevation and angle spot-on. There's a large brace/jack on the outside of the rails that was aligned first. This unit uses friction to grip all four rails and, when air pressure is applied, pneumatically draws them closer together.


DSC02084 DSC02090 DSC02092

The workman was constantly measuring the gap between the rails until it was within tolerance. I note that they didn't just bring them closer in one hit. The workman operating the compressor only applied pressure for short periods, maybe 3 seconds each time, and then his colleague would measure again. They'd then let the whole contraption rest for 5 seconds before applying further pressure.

I'm actually astonished that the rails even moved. It's a relatively straight section of track and where does the slack come from? I can't imagine they actually stretch the steel? They had removed around 6 rows of clips from the original rails to allow freedom of movement. I'd say the gap was about 40mm to start with, but they brought it right down to less than 20mm!

Once it was as close as required, the rails were further adjusted using (what looked like) rail spikes between the closest sleepers and the base of the rail. These were hammered in to raise or lower the rails. The worksman spent around 5 minutes doing this, making sure everything was totally level. The ruler was around a metre long, allowing him to see how much the rails tapered off on either side. You could tell he did not want to create any kind of gradient on either side of the join.

Step 2 - Build The Mould

From what I could see, the mould was made of some form of terracotta? The colour was the standard red, but it really could be made of any sort of compound. The base was removed from the packaging first and a layer of glue applied down each long edge. This was then smoothed around the edge and along the ridge to make sure that there'd be no gaps once joined. This base was then placed into the metal tray which would soon form the case that holds the mould together.

DSC02094 DSC02095 DSC02098

The whole base was then placed under the gap between the rails. It was held in place with clamps and, once again, adjusted once and again to make sure that it was completely square and level with the join.

DSC02096 DSC02100 DSC02101


The sides of the mould were then taken out of the pack. These were placed in the 'sides' of the metal case and both units were then assembled onto the rail. From here, the mould was complete, with an opening at the top where molten metal would be poured in? You could see that, at the top of the mould there was one short edge that was lower than the others; presumably this was for excess molten metal to flow over.


Once this was done, the final outer frame was dropped on and the 'drip tray' inserted on the side that had the overflow 'spout'.


DSC02107 DSC02108 DSC02111

From here, a large block of malleable clay (or other playdough-esque material) was split up and used to totally fill all gaps around the edges of the case and mould. The goal was to ensure that all heat, once the mould was full, was to stay trapped inside.


The entire block was applied to all facades of the mould, specifically where the case and the mould met.

Step 3 - Add Fire!

A cute little terracotta (or other substance) pot had been sitting on the back of the work truck for a while, but it was now its time for action.


To 'set the mould', a small mount was clamped to the rail and then the oxy-acetylene torch was flamed up and positioned on it, aiming the fire directly into the mould. The flame from the torch burned a strong blue, but eventually yellow flames began rising from the openings on either side of the torch. One of the worksman then grabbed the terracotta pot whilst the other pulled the torch out of the mould. Without being able to see inside the mould, one couldn't really work out what exactly was being heated, but you'd have to assume the rail ends were red-hot by this point!


DSC02122 DSC02123 DSC02124

The pot was placed directly on top of the whole mould and the torch was placed in the small opening on top.


I'll let the following video explain the rest...

A lot of trust placed in a set of serious gloves!

Step 4 - Clean Up

From here, there was a 10 minute break whilst the burning-box-of-death cooled down. The overflow tray on the side was thrown track-side once it could be lifted.

DSC02132 DSC02130 DSC02134

After a while, the outside metal case was unscrewed and whacked a few times until it fell loose. Now the mould was perfectly visible, and perfectly-solidly-formed as a single unit, welded to the rails!


To get this off, another utensil was used. The workmen brought over another flat-frame-style machine and placed it over the mould. With a few clamps, they secured it on all corners to the rail and then one of the guys started pumping a lever handle. I couldn't quite see what the action resulted in from where I was, but I assume there was a flat blade that was slowly, flush with the railhead, cutting into the mould on a horizontal plane.


DSC02143 DSC02144 DSC02147

The other worksmen started beating the mould with a mallet when the cutter wasn't being used. After a lot of intimidation, the mould finally started budging... but instead of a clean splice, it decided to split to pieces... pieces of 1000 degree red-hot danger.


More time was spent letting the bits cool down and then it was all moved to the rubbish pile track-side.


From here, a portable grinder-on-wheels was used to finally trim down the weld. Once complete, it finally resembled rail!

I'd actually watched the 2nd of 4 joins be welded before I had decided to get my camera and record the one above. This part, the removing of the mould, was much smoother on that one; a single knock after a slight clamp saw the whole lot just break free in one hit. There was no grinding required afterwards either! I hadn't really noticed anything done differently between each session, so I wonder how many variables come into play when it comes to doing this and how easy it is to stuff up!?

Gantry Foundations

Whilst all the welding was taking place, another vehicle had been busy drilling a column on the other side of the track. The colour of the earth was quite interesting, being somewhere between clay and red earth? A re-bar metal frame was then built up and inserted, with the square frame of pre-built bolt rigging for the base of a pylon. The alignment of this frame in the hole was actually a big thing and watching the surveyor get it correct was pretty interesting.


Turns out there was a remote surveyors camera sitting half way down the track, fixed on the location of the pylon.


And, with a remote mirror, with a very fine tip, the surveyor measured each corner of the frame. The workers around nailed, cut and hammered the external wooden frame to get the metal frame in the exactly correct position.

DSC02161 DSC02162 DSC02165

The surveyor was holding a handheld computer that was relaying the stats from the surveying equipment.


Good to see technology helping all departments.

Tamping And Cleaning

Whilst everything else was going on, there were also tampers and ballast cleaners working away. Turns out they were tidying up a cut-in insulated joint that had been installed the night before.


DSC02167 DSC02082 DSC02169

The ballast cleaner sounded like it was in pain, chewing rocks up and spitting out a lot of dust.


After all of the above, they managed to clean up and the trains were running again the next morning.

Filed under: AUS No Comments