Did you hear about the Christmas Eve train crash near Oxford? It occurred at Shipton-on-Cherwell to be exact. Thirty four people died and sixty nine were injured. You didn’t? Well, you do have an excuse, as it happened in 1874. It was the worst disaster to befall the Great Western Railway and the cause was a broken tyre on the leading carriage. It initiated reforms that were pivotal in the advancement of railway safety, particularly in wheelset design.

This was how Railway Consultant Adrian Shooter commenced his keynote address at a seminar held on 30 May at the Institution of Mechanical Engineers in London. Entitled “Axles, Wheelsets and Bearings – Balancing Safety, Performance and Cost” the IMechE event hosted presentations from eminent engineers, consultants and industry experts.

So are we safe?

It has been 16 years since a defective wheelset caused a major accident in the UK. At Rickerscote near Stafford on 8 March 1996 a freight train was derailed because of a broken axle on a tank wagon. A postal train hauled by a Class 86 locomotive ran into it.

A mail sorter was killed and 22 others were injured, including the driver of the locomotive, which came to rest against the end wall of a house.

Wheelsets are some of the most expensive consumables on rail vehicles. They account for a significant proportion of an operator’s maintenance budget and are critical to vehicle availability. They can also be key components in causing damage to infrastructure. The rail industry is under great pressure to reduce costs and increase its value to customers, but this must be balanced against operational performance and above all, safety.

It is estimated that there are in excess of 144,000 wheelsets in use in the UK and they must all be maintained, inspected and tested to exacting standards. As Adrian Shooter put it, “We must be assiduous in making sure that the right standards are adhered to. Procurement is driven by cost, but only by strict adherence to standards can it be ensured that we can all go about our business safely.”

Safety management

So what is the risk to be managed? Cliff Cork, head of infrastructure and rolling stock at the RSSB, posed this question and went on to discuss the applicable Standards framework and the ways in which European legislation impacts upon it. Risk and harm are measured in terms of fatalities and weighted injuries (FWI).

The short answer to the above question is that the risk attributable to rolling stock is 2.2 FWI/year (out of a total of 140.9 FWI/year). Of this, axle and bearing risk accounts for 0.64 FWI/year and wheels 0.066 FWI/year. In other words, the risk from wheelsets can be calculated to cause 13 deaths in 100 years. This amounts to 6% of rolling stock risk.

Cliff said: “We would all agree that the low level of risk that such a high population of wheelsets impose is due to having an agreed set of rules or standards in place, which we all follow.” The risk may be small, but the consequences of getting things wrong can be disastrous.

Within the UK, wheelset and bearing design is governed by Technical Specifications for Interoperability (TSIs) with supporting standards, typically EuroNorms (EN), and Railway Group Standards. The essential requirements are safety, reliability and availability, environmental protection and technical compatibility.

Setting standards

Roger Deuce, senior project engineer, bogies at Bombardier, also began his presentation with a question: “When is a hole not a hole?” To illustrate how EN standards definitions are reached he described how an axle that has a central hole to facilitate ultrasonic testing may be defined as ‘solid’.

The axle is ‘solid’ if the journal diameter is more than three times the bore diameter and the wheel seat diameter is more than four times the bore diameter. EN wheelset standards are far reaching and include both design standards and product and process standards. They govern such things as material grades, permissible stresses, testing and batch testing, vacuum degassing, residual stress surveys, surface finish and fatigue testing.

A typical EN process related wheelset fatigue test will involve resonance testing in a jig. The axle is mounted vertically, clamped at its lower end and fitted with a motor driven mass imbalance at the upper end.

The introduction of EN wheelset standards is seen as a significant step forward. But, as Roger Deuce pointed out, “They are not a substitute for good engineering practice and the experience of a competent design authority. EN standards have helped to improve standardisation across Europe, although there are still operator/network variations.”

Shared system

At the sharp end of railway operation, it is incumbent upon every operator to ensure that its vehicles are maintained adequately. The roles and responsibilities in a shared European system were outlined by Richard Lockett, head of Cross Acceptance Unit, European Railway Agency. Based in France, the ERA has responsibility for enhancing the level of interoperability of railway systems and developing a common approach to safety on the European railway system. Member states must ensure that railway safety is generally maintained and, where reasonably practicable, continuously improved.

The use of vehicles (such as wagons) and wheelsets may be shared between different operators. The safety management systems of all the users must manage this, but as Richard Lockett pointed out: “The management tools previously employed by BR, SNCF, DB, etc to manage their own single actor systems are not fit for purpose for a shared system.

Old and inappropriate roles and responsibilities remain, such as Network Rail ‘approving’ private wagons and safety authorities specifying maintenance schedules and testing procedures.” The use of shared components requires complex contracts agreed by many parties. Maintenance, checking and overhaul specs are also a compromise and again many parties need to agree. Across Europe, conformity with EN standards is voluntary, whereas in the UK we have Railway Group Standards. Richard Lockett asks: “How can a RGS fit in where TSI’s supported by EN’s cover the whole network?”

Perhaps lessons can be learned from the USA where there are many operators, many owners and many repairers. There are however only five designs of wagon wheelset and one safety authority. In Europe there are thousands of wheelset designs and 25 safety authorities.

Roger Lockett summed up: “Wheelsets are a technical challenge. Maybe diversity of design and incomplete implementation of the directives roles and responsibilities makes the logistics and management control an even bigger challenge!”


Dr Alan Lawton is an independent engineering consultant. His research into wheel lathe best practice has revealed large differences in the average depth of cut between lathe operators. He also discovered significant variations between depots, with Slade Green making lighter cuts than other depots in the study. He explained: “Within each depot, if all operators could do what the best operator does, it would give another ‘turn’ for each wheelset, i.e. 25% longer wheel life.” Furthermore, “If each depot were able to exploit the same ‘light cut’ that works for Slade Green, there would be an extra two ‘turns’ on each wheelset!”

That would equate to 45% longer wheel life. Alan suggested this could be achieved through closer management and support for lathe operators by analysing wheel turn records and trials of lower cut depths.

On trial

Wheel tyres of ‘Superlos’ steel, manufactured by Lucchini in Italy, offer significant improvements in wheel life. A trial has been undertaken on Alstom Class 175 DMUs, which normally have their wheels turned at approximately 75,000 miles to eliminate surface cracks caused by Rolling Contact Fatigue (RCF). Alstom removes 5mm radius as standard, allowing six cuts per wheel over its lifetime, i.e. 525,000 miles in total. The Superlos wheelsets have run over 200,000 miles before visiting the wheel lathe, giving an expected total mileage of 1,000,000 miles. The only problem is that no one at present understands how Superlos inhibits RCF cracking!

Axle Safety

There were 37 UK in-service axle failures between 1975 and 2002. In the final ten years of this period there were only nine failures, and that trend continues. Fatigue cracks initiated by surface defects are the main cause of cracking, with corrosion being the chief culprit. Alan Lawton explained how axle design standards assume an infinite fatigue life and presume that axles retain a smooth round surface. “They cannot deal with surface damage or variations in maintenance and inspection,” he said.

“Nor can they identify appropriate processes for axle maintenance and inspection. There is no knowledge base from which to identify sound, logical maintenance and inspection processes for axles.” It is recognised that money is wasted looking after axles, although this ensures of course that we stay safe.

Alan Lawton suggests that we need to understand real axle loads, how corrosion assists fatigue cracks to initiate, how effective non-destructive testing (NDT) really is, and how axle surfaces become damaged. This would help us to identify more economic maintenance and inspection rules, without compromising safety.

To this end, the RSSB has been working with DeltaRail to understand real axle loads on Class 319 EMUs and on Mk IV coaches. At the same time, innovative work from Prof. Stefano Beretta, Politecnico di Milano, has provided new understanding of the initiation and growth of corrosion assisted fatigue cracks in axles.

The RSSB has partly funded an EU research project known as WIDEM, which aims to optimise the design and maintenance of wheelsets, reduce whole life costs, reduce un-sprung masses and extend maintenance intervals. This work could in time provide the engineering science to justify elimination of depot NDT, the elimination of axle painting and the introduction of axle skimming.

Smaller wheels, higher loads

UK and UIC standards limit the ratio of wheel load to wheel diameter (Q/d) to around 0.13. Alan Lawton explained that this limit can be linked back to 19th century calculations of “Hertzian contact stress”. RSSB Project “T889 Q/d” attempts to establish the current population of Q/d values and calculate the distribution of Hertzian contact stress on the railway. It also uses Hertzian stress based computer models of rail damage to assess the effects of increased Q/d.

Questions remain however, such as the real nature of stresses in wheels and rails and how damage can accumulate.

Alan Lawson thinks we need to understand better the mechanism for the initiation and growth of damage in different wheel and rail steels, but he admits there is no real business case for the research. But does this mean it shouldn’t be done.

Predictions of how wheel and rail damage might grow could bring benefits such as 30 ton axle-loads, commonplace in the USA.


Funded by RSSB and WOLAXIM (a collaboration between EU SMEs and research organisations to develop novel NDT solutions for the assessment of railway axles), the Politecnico di Milano, under Prof. Stefano Beretta, has attempted to produce a corrosion fatigue model for a railway axle and assess its impact on safety. Results have shown a continuous decrease of fatigue life as corrosion continues.Testing involved dropping artificial rain (pH6) onto test axles rotating under load. Surface cracks were observed to grow from corrosion pits, eventually coalescing and deepening.

The phenomenon is known as Hoddinott cracking, after Dudley S. Hoddinott, an HMRI Inspector who first observed the phenomenon during an investigation into a derailment at Shields Junction. It is characterised by a surface pattern of micro cracks, all lying in roughly one direction. The effect is worst in transition areas, e.g. where the axle meets the wheel seat.

The research concluded that crack growth rate under corrosion-fatigue can be described with a simple mathematical model. This enables a ‘nucleation lifetime’ to be calculated for an axle under given service conditions. An argument follows from this for not painting axles and instead applying a simple surface grinding process in axle transition areas.


A large number of axles are scrapped every year due to corrosion. Pools of spare axles are therefore integral to the smooth running of the industry, but industry fragmentation resulted in multiple part numbers, drawings and numerous pools of axles for different fleets.

David Wilson of Porterbrook described how rationalisation has now allowed a common pool to be established. Permitted corrosion limits are strict, hence the high scrap rates, but there has been little documented guidance on wheelset reclamation. David Wilson would like to see an agreed approach in developing procedures for the removal of corrosion and light damage. Removal of material will have structural implications, but of what significance? Studies on Mk IV wheelsets have shown that reclamation is possible, but there are limitations (on axle loading) and it must be controlled.

Common ROSCO (ROlling Stock Leasing COmpany ) procedures have now been drawn up and packaged into a useable specification, giving guidance on inspections, repairs and branding of axles. Ultimately this should bring significant cost benefits to the industry.

Out on the real railway, pragmatic problem solving is the order of the day. In the second part of this article we will see how manufacturers, maintainers and TOCs have already excelled in the balancing wheelset safety, performance and cost.