On 17 October 2000, a train derailed at Hatfield, Hertfordshire when the running rail catastrophically fractured caused by rolling contact fatigue (RCF). Four people were killed and more than 70 injured.
If good can come from such tragic events, it is that the lessons are learned and applied to improve safety. The Vehicle/Track System Interface Committee (V/T SIC) is a product of the Hatfield tragedy. The seeds of cooperative research and development were sown with the creation of the independent Wheel Rail Interface Systems Authority in May 2001, which morphed into V/T SIC, sponsored by RSSB, in 2004. It was, and still is chaired by Network Rail’s Andy Doherty.
In the early days, there was a view that “the trains were causing all this RCF”. At one level, of course, this was true – if no trains ran there would be no RCF – but it didn’t help solve the problem. Eventually independent engineers started to explain the engineering reasons for RCF, which was a precursor for solutions.
This started the process that V/T SIC has employed very successfully; carry out enough research and development to be confident an improvement can be made, give advice for implementing the findings, whilst continuing R&D to further improve knowledge. The introduction of grinding as a preventive maintenance/life extension activity was one of WRISA and V/T SIC’s early successes.
Each year, there is a seminar to report on progress made with the various programmes being managed. The 2017 seminar was held on 4 December and was attended by over 70 delegates from across the industry.
Andy Doherty introduced the session by reviewing what has been achieved since the Hatfield crash. There is strong competence in the wheel rail interface across the industry, with clear and competent standards for track and rolling stock asset management. There has been a reduction in fundamental energy and damage in the wheel-rail interface (WRI) leading to a reduction in defects in both rail and wheel leading to a significant reduction in broken rails.
Engineers now have the ability to model the WRI using tools such as Track-Ex (identify risk of RCF and propose corrective action) and the Vehicle Track Interaction Strategic Model that enables engineers to be ahead of the defects. The variable track access charge has provided the financial incentive to the industry
These are good achievements, but there is more to do, particularly as traffic is growing and slots to access the railway are shrinking. Andy’s target list for further improvement included:
- Rail Management;
- Reduce broken rails below 50 per annum;
- Reduction in squats and reduced maintenance costs;
- Managing out rail foot corrosion defects;
- Adhesion and autumn;
- Braking systems that do not rely entirely on adhesion such as magnetic track brakes (MTB);
- Fully integrated brake control including electric brake and MTB;
- Wheelset Management
- Further wheel life extension;
- Additive wheel repair;
- Reduction in whole life cost;
- Truly low-cost track for regional and light railways.
At this year’s seminar, Andy delivered the first talk himself, about the opportunities to obtain funding for research under the Horizon 2020 and Shift2Rail EU programmes. He reminded everyone that UK had been a powerful influence in structuring Shift2Rail using the Rail Technical Strategy (RTS) to help set the agenda.
Much good work has been done and, for 2018, Network Rail has €9 million to spend and is working with its supply base to try innovations with a high Technical Readiness Level. This is part of a €30 million programme being led by Network Rail, and Andy highlighted several initiatives: a scanning tool to “look” behind tunnel linings, the ability to lift brick-arch bridges (issue 146, December 2016), induction welding and the development of an automated squat defect repair method.
Andy then introduced the remainder of the seminar, which was a mixture of progress reports on research projects and feedback from the implementation of previous research.
Dr Gareth Tucker from Huddersfield University sought to answer three questions. Can flange height and thickness limits be changed to allow more wear? Is there a benefit? And what additional controls might be required if flange wear limits relaxed?
Gareth used the wear limits defined in Railway Group Standard GMRT2466 on Railway Wheelsets, compared them with those in the Locomotive & Passenger and the Wagon TSIs (Technical Specifications for Interoperability), and sought to understand what hazards are controlled by flange wear limits such as negotiation of obtuse crossings.
GMRT 2466 defines nine wheel profiles with minimum flange thickness of 21mm (P9 profile) to 27 mm (P5, P10). Maximum flange height varies between 33mm (P5, P10) and 36.5mm (all except P5, P10, P11).
Gareth analysed the impact if all profiles were to have a minimum width 24mm and maximum height 36.5mm by assessing, or modelling, the performance of various thicknesses against the following hazards:
- Flange back strikes open switch;
- Flange back makes contact with check rail before checkrail entry point;
- Flange back makes contact with wing rail before wing rail entry point;
- Flange hits the end of a closed switch;
- Flange makes contact with a closed worn switch blade with a low contact angle (relative to horizontal);
- Wheel with high tread wear makes contact with a fish plate on rail at maximum head wear allowance;
- High angle of attack in curves due to thin flanges;
- Allowing wheels to run longer between turning cycles could lead to low or high conicity;
- Flange strikes obtuse crossing nose.
The results to date indicate that the flange height limit for P5, P10 and P11 could be increased to 36.5 mm and, pending review of worn switch monitoring, it may be possible to introduce 24 mm flange width for P5 and P10. Gareth also highlighted that TSI flange width limits should only be used in combination with full TSI wheelset dimension controls.
Gareth noted that allowing thinner flanges might only be of any value on the final “life” of the wheel before scrapping because allowing wheels to wear to the thin flange limit generally means more metal has to be turned off to restore the correct profile, thus reducing overall life.
Brian Whitney, Network Rail’s Engineering Expert (Track and Lineside) gave his annual update on rail management – he presents the trend in broken rails for Network Rail each year.
In the 40 years from 1962 to 2002 there was an average of 750 broken rails a year, with an absolute peak in 1980. In 2000 there were 919, but the number has declined rapidly with less than 100 forecast for the whole of 2017 – ten of which occurred between Ferrybridge and Llangerrech on 30 October 2017 following the passage of a heavily laden freight train with very severe flats on one axle!
This reduction is all the more remarkable given that tonnage has increased by approximately 50 per cent and access time to conduct repairs and maintenance has been reduced.
Brian’s work is now focussed on identifying the pre-cursors of broken rail to give track maintainers the tools to identify track defects that, if ignored, might lead to a rail break. This is easy to say, but to identify individual unsupported sleepers, small plain line non-actionable dip angles (say, 10 to 15 milliradians) or wet spots (especially at bridge transition) on rails with high cumulative tonnage is not easy. Brian illustrated these factors with examples from previous broken rail investigations.
The challenge is to find a tool that can identify wet spots from aerial photographs, pick up the small plain-line non-actionable dip angles and integrate this with the cumulative tonnage, age of rail, supervisors’ inspections to deliver a rail health index/risk ranking.
Brian also talked about the benefits of tools that will help to identify pre-cursers, including Sperry B-Scan, train mounted eddy current testing and fitting deflection sensors to the measurement train on both the locos and the passenger cars to understand whether different loads cause variation in deflection and hence point to inadequate support.
Whilst good progress has been made on rail management, more needs to be done to design out switch and crossing (S&C) failure modes and to detect faults before failure. Brian made the rather pointed remark that the same S&C designs have been used for over 50 years.
He concluded his talk by saying that tools are increasingly available to improve decision-making. His task is to provide guidance on their use so that:
- Wider use is made of premium rail steels in key locations;
- Tensile residual stress in the rail foot is reduced to minimise foot defects and breaks;
- Rail grinding and milling are used optimally to reduce surface damage and the need for extensive re-railing;
- Pre-cursor conditions are proactively sought out; today’s geometry fault maybe tomorrow’s rail defect or break;
- Design cant and cant deficiency in curves for the dominant traffic is considered – high tonnage low speed freight vs low tonnage higher speed passenger service;
- Track support – stiffness and resilience – is maintained and improved ways of repairing the track support are developed.
Line speed differentials
Mark Burstow, principal vehicle track dynamics engineer at Network Rail, gave two presentations. The first concerned line speed improvements. For many years, ‘Sprinter’ vehicles have been permitted to operate at higher speeds than other traffic on some secondary routes. Over the years there had been some ‘drift’ in what constituted a Sprinter, and research project T996 recommended characteristics to define Sprinters. The work did not address the reasons for Sprinter differentials.
V/T SIC requested a “bottom-up approach to determine the design speed profile for the current infrastructure”, or, as Mark put it, to determine if current differentials on infrastructure are appropriate. The original reason for the application of differential is often not clear and, following renewals or enhancement, the original reason for the differential may no longer exist. Indeed, it might be acceptable to raise line speed to the Sprinter speed for all traffic in some cases.
The York – Scarborough line (YSL), which has Sprinter differentials applied, was chosen for the study. It is a candidate for operating a wider range of vehicles whilst maintaining or improving current timings; for example, Class 185 trains (not classed as Sprinters) which are currently restricted to lower speeds.
The factors that impact on line speed, and therefore need to be assessed, are curvature, switches and crossings, level crossings, structures and signals (spacing and sighting). Mark said that jointed track would lead to restricted speed, but most of this had been replaced by CWR (continuous welded rail) since the differential speed was first adopted.
Mark explained that the assessment had used the principles in Network Rail’s standard for Design and Construction of Track, gauging, structures’ RA ratings and the output of T996, and had assumed that the track category would not be altered. Work undertaken to date suggests that there may be a good case for raising line speeds.
Premium rail grade selection
In his second presentation, Mark Burstow provided practical advice and guidance to track engineers and designers on when to use the various grade of rail available. He said that current guidance is only slightly helpful: “Premium hardened steel may be used on one or both rails where rail life is reduced due to surface damage such as RCF or there is a high rate of side wear, a significant flattening of the low rail, or corrugation” (NR/L2/TRK/2102).
This advice pre-dates the most recent developments in rail steels and before the development of the tools now used to predict rail damage. Mark presented evidence to show that harder rail is not a panacea for all ills. He showed a number of scenarios where the use of premium grade rail will indeed reduce wear but will also increase RCF.
Mark’s preference is for wear rather than RCF, as the former can be more easily managed, for example by lubrication. In summary, adopting current grades of premium steels for all curves of radius less than 2,500 metres is not optimal; it is of little benefit on curves with cant deficiency and there is significant disbenefit on tight radius curves.
There were also presentations about the RSSB-sponsored enhanced sanding trials given by Steve Mills (issue 157, November 2017) and the development of rail steels from Jay Jaiswal and Adam Bevan of the University of Huddersfield.
Following a final panel session, chaired by Virgin Trains’ Keith Mack, Andy Doherty gave some final thoughts on the work of the V/T SIC, thanked everyone for attending and looked forward to the next meeting.
V/T SIC exists to: Define the geometry condition of the wheel/rail interface; Set out permitted track geometry, including curve radius, installed cant, vertical curvature for the various types or track; Set out allowable wheelset parameters, such as gauge, conicity, profile, axle load, for the trains that operate on the network; Set requirements to maintain the wheel/rail interface within the prescribed limits and ensure safety of the railway; Implement processes to ensure sustainability of the railway; Ensure systems exist for seasonal maintenance activities and revised to account for seasonal variation; Utilise systems to monitor the condition at the wheel/rail interface and to analyse and distribute data collected by these systems to the appropriate party for the monitoring of performance and programme rectification activities. It currently has three sub groups: Adhesion Research Group; Vehicle/Track Technical Advisory Group; Wheelset Management Group; Plus an infrastructure sub-group to be added in 2018.
This article was written by Malcolm Dobell.
Read more: VolkerFitzpatrick – Upgrading stations