HomeInfrastructureBroken rails - causes and prevention

Broken rails – causes and prevention

Listen to this article

“Let me paint you a picture”. That was how Marc Clarke, technical lead for rail welding with London Underground, started his presentation at the twenty-fifth technical seminar of the Institute of Rail Welding recently. He accompanied this with a picture of a welder carrying out a repair on a flat bottom rail on a section of overground track.

Having seen rail repair welding first hand, the first thought that came to my mind was – welding parameters. Currents, preheats, weld bead lengths and how key it is to get these right to produce a sound weld.

“Great,” I thought. “I might learn something new about rail repair welding as it is something I am involved with.”

However, by the end of his presentation, I can honestly say I had learned nothing about rail repair welding in terms of technique. Marc, though, was true to his presentation title – “Challenges of Welding on the London Underground” as he gave us a vivid description of what it is like. I have to admit that I had never considered the non-technique challenges of welding underground but, from what was described, it sounded as though it can be, on occasion, a bit of a logistical nightmare.

Before any welding can begin, the welding team needs to get its equipment to the required underground location. Access and egress points? Not quite, more like up to 60 metres underground down steep public escalators carrying welding boxes, torches, gas cylinders, PPE and lighting.

And the rail? Well, that has to be shunted in from the nearest overground connecting tunnel which can be half a kilometre away.

Then there’s the issue of parking the van two blocks away from the station (if you’re lucky), hassle from the ticket office getting in and out of the station, permits to work – and all this before any welding has commenced.

“Welding underground in the summer is like being in a pressure cooker, the conditions are horrible,” Marc continued.

We have to remember that the London Underground is over 150 years old, making it the oldest system in the world. Consequently, it is inevitable that
a variety of building structures will be present, spanning the last 150 years. Unfortunately, this includes asbestos sites where it is mandatory to wear breathing apparatus – in the already stifling conditions.

So the difference between welding overground compared to underground can sometimes literally be like day and night. It’s no wonder Marc says that they have some of the best welders in the world working for them. They have to be, to be able to adapt to these difficult working conditions.

However London Underground is moving with the times and looking to make improvements to how it works from many angles. One of these is the approval of premium rail HP335 for use on the network. This will reduce the frequency of work required in terms of rail replacement and weld repair, and ultimately reduce life cycle costs.


Standards galore

After Marc’s vivid oil painting, Brian Whitney, principal track engineer for Network Rail, gave us some insight into three aspects of his job role
– standards, standards and standards. Brian put up pages and pages of standards, all of which needed updating and revising, a process with which he is heavily involved.

This is another job I wouldn’t fancy. I’m not sure what would be worse, welding on the London Underground or being responsible for making sure all the standards that Network Rail staff work to are bang up to date (not to mention all the other UK rail operators, maintainers and installers that use these as the benchmark).

Considerable effort is invested in making standards easier to reference, to include learning from experience, research and trials. Standards also need to evolve to take into account new inspection technologies and frequencies, for example the work with Bob Crocker and the eddy-current trains – more on this later.

One of the positive outcomes that can be linked to improving standards is the decreasing number of rail breaks. In 1998, there were 952 rail breaks, and from the chart Brian put up, the number of rail breaks has been decreasing each year down to 109 in 2015. As you can imagine, the number of high-risk rails found before failure is sensitively linked to the inspection frequencies dictated in the standards. Overall, that is around a 90 per cent reduction in rail breaks and that, against the backdrop of heavier axle loads and 50 per cent increase in traffic, is something to be quite proud of, I think.

Brian talked us through the individual changes that have been implemented and which have contributed to the decreases in rail breaks over the past 15 years. These included a campaign of re-railing, grinding strategies, ultrasonic inspections and the tightening up of dip angles.

Dip angle limits have been improved twice at different times, and each adjustment is thought to have led to a decrease in rail breaks. As Brian explained, it is becoming clear that even small plain line geometry faults, which are well within their limits, are correlating with rail breaks. Dip angles, changes in track stiffness at discontinuities such as bridges, crossings, areas of ballast settlement – all are coinciding with rail breaks. However, the number of variables that differ at each site is significant and, as yet, clear correlations have not been determined. Watch this space.

A handy tip to spot ballast movement, which can undermine the sleepers and rail, is to look for whitening of the ballast, a function of friction acting when the ballast is moving. Normally it is very difficult to see ballast settlement when walking the track; it is best done from some height using aerial photography or a drone.

Geological features, that affect track bed stiffness, differ from one part of the country to the next, making this a complex challenge. “We are accustomed to working on the top half metre, with minimal interaction below this, but this is something we need to tackle moving forward,” Brian commented. It certainly is a big challenge considering that our history of line formation is based on building embankments from cuttings.

The statistics presented are only as good as the method of recording data in the Rail Defect Management System (RDMS), which is also evolving. A good understanding of the limitations and nuances of the data recorded in systems and registers allows the user a deeper appreciation of the state of the network, and also allows a comparison of statistics with other rail networks.

The French connection

Brian has been working with his counterparts in France’s SNCF. It has taken some time for the relationship to get to where it is and only now is SNCF starting to open up a bit and appreciate the benefit of collaborative working. Our network is similar in size to SNCF, so it makes sense to share notes, “but we need to make sure we are comparing apples with apples”. Subtle differences in the reporting and categorising of rail breaks can make all the difference, so it is vital to understand these in order for a meaningful comparison to be made.


A lot of critical thinking and judgement is required to assess numbers and statistics. For example, squats come out as the top defect. But is that really because there are five times more of these than any other defect, or is it just because squats are more easily detected?

Interestingly, SNCF is considering specifying softer rail grades such as R200, believing in the benefit of using rail with a high wear rate so defects do not have time to develop due to material loss. In the UK, our experience is taking us down the route of using HP335 and harder rail grades.

How do we reconcile this huge difference in strategy? As Brian explained, people need to appreciate that the UK network is operated in a vastly different manner to that in France where no-one would dream of putting freight on a passenger line as we do; they have separate tracks for each which means that the strategies for each network necessarily need to be different.

Assessing skills

Continuing the theme of standards and regulations, Paul Flynn, training evaluation and assurance team leader, talked about industry standards, and how to be compliant. Paul presented the new Skills Assessment Scheme, which is a method of ensuring that process practitioners such as welders are continually assessed to ensure that they have had the necessary development, post training development, interim checks and renewal checks.

This scheme utilises a competency-based approach rather than the main focus being a ‘one size fits all’ knowledge test as was the case previously. In addition, the requirements for each process now reflect the risk involved, resulting in a more proportionate approach rather than a generic system.

As briefly mentioned earlier, Bob Crocker has been working on the eddy-current inspection train, the proper technical term being RSU (roller search unit). Bob works for Sperry Rail, and he explained that trains with eddy-current sensors mounted on them are continuously running around the network, mapping out where RCF (rolling contact fatigue) is present and categorising its severity. These trains cover the entire rail network every eight weeks.

Sperry Rail has devised a system that utilises sensors at 70, 37 and 0 degree angles which move along with the contours of the rail, providing more complete coverage of the head.

RCF is a by-product of steel-on-steel contact at the wheel-rail interface; small cracks that can grow with time and can lead to a transverse rail break with catastrophic consequences, such as Hatfield. Hence RCF needs to be managed and the rail ground before the cracks can grow or replaced entirely if the RCF cracks are too severe.

Historically, the depth of RCF-generated sub-surface cracks, the type that can lead to a transverse rail break, have been correlated to the crack length present on the surface of the rail. Bob says this is a loose relationship, and it means that railway administrations may have to spend more money managing RCF at certain locations than may be strictly necessary due to the fear of the potentially disastrous consequences.

So what has Sperry Rail found? “RCF is everywhere, not just the high rail and fast lines, but tangent track and S&C, it was a revelation!”

Sperry Rail reports back to Network Rail which decides what the best course of action is – grinding or re-railing. Following successful trials, milling is something that may be introduced as an intermediate between the two. Rail milling removes much more material than grinding and essentially allows Network Rail to ‘reset the clock’ in terms of rail management.

Similar inspections have been carried out in Holland and Belgium, and Bob commented that everyone has been amazed by how much RCF is present, in particular at unexpected locations.


Another method of reducing rail breaks, and increasing the longevity of rail in corrosive environments, is to make use of British Steel’s rail protection system called Zinoco, which Sean Gleeson explained.

Zinoco was developed following years of development work in the laboratories, followed by trial installations in the UK and abroad which are performing very well, indeed better than any other coating system available. Zinoco stands for Zinc-no-corrosion. “The four nasties, coastal routes, level crossings, third rail and tunnels, are locations where rail corrosion is most prevalent,” Sean told his audience.

There are two methods for corrosion protection, a barrier or sacrificial. Zinoco provides both an excellent effective physical and electrochemical protection. For example, epoxy coatings is good but, if it is breached, then corrosion will be focused at that point and will eat away the steel below the barrier.

“Railcote, British Steel’s previous rail corrosion system, is good but its physical durability has limitations” Sean stated. Railcote comes in 18-metre lengths; longer ones are impractical due to delivery damage that can occur.

This was the driving force for the development of Zinoco as Network Rail required longer lengths. Zinoco is available in 108 or even 216 metre welded lengths. British Steel selected the final Zinoco product based upon both its corrosion performance, mechanical resistance to damage such as impact and abrasion resistance and its damage tolerance. Zinoco’s superior performance can be clearly seen when placed side by side with other coatings that have all been scratched.

British Steel has a long history researching and supplying corrosion protection. Zinoco’s enhanced durability and corrosion protection will play its part in decreasing the number of rail breaks in years to come.

The range of speakers at this year’s Institute of Rail Welding Technical Seminar certainly managed to keep a knowledgeable audience interested throughout the day. From the trials and tribulations of welding in confined spaces, through tackling the ever-present problem of rolling contact fatigue, to preventing rail failure through corrosion, the topics were varied and informative.

I hope to be lucky enough to be asked to go again next year…

Written by Dr Qasam Javaid, rail technologies consultant with British Steel.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.