Understanding the condition of substructure is vital for maintaining good track geometry. With trackbed investigation now recognised as good practice, some engineers might dismiss this statement from the 2012 Rail Technical Strategy (RTS) as obvious. However, the issue is how to obtain this understanding with increasingly limited possession times for trackbed surveys. So it’s hard to disagree with the RTS’s call for improved trackbed assessment techniques.

One company developing such techniques is URS, which has a single supplier framework contract with Network Rail for trackbed investigation and design. Its trackbed investigation work has developed techniques that are now industry standard practice. This has also resulted in two innovation awards. In November 2010 URS’s “Total Route Evaluation” techniques won British Expertise’s Innovation award. In March its Rail Trackbed Stiffness Tester was a joint winner of the innovation award at the Railway Industry Association (RIA)’s Technology and Innovation Conference as reported in last month’s the rail engineer.

Total route evaluation

In 2001, a report to the Rail Regulator showed that Railtrack was spending £200 million per annum on track renewals yet not achieving the required track quality improvements. At the time, trackbed investigation was traditionally undertaken by slit trenches up to 200 metres apart. This was slow and labour intensive and did not identify issues between sample points and beneath the ballast, which often led to incorrect track renewals specification. Such investigations are essentially required to assess trackbed condition and track layer stiffness, the elastic rail deflection under wheel loading. The problem is that the subgrade, buried under the track, is the primary determinant of overall track stiffness. Hence there is a clear relationship between trackbed stiffness and track quality.

In addition to the requirement to optimise track renewal specifications, an understanding of the subgrade is needed to determine the track’s critical velocity. This is the speed at which excessive ground vibrations occur as the waves from trains’ shock loading on individual sleepers resonate with the ground’s natural frequency. Typically, this starts to be an issue on soft subgrade at speeds of 140 km/h.

So it is not surprising that trackbed investigation is the subject of various research projects. For its part, Scott Wilson (since acquired by URS) started trackbed research 15 years ago through a joint venture with the University of Nottingham that became URS’s Pavement, Trackbed and Materials Consultancy. This led to the development of the Total Route Evaluation (TRE) methodology which uses a number of techniques including Automatic Ballast Sampling (ABS), Ground Penetrating Radar (GPR), Springbox testing and Falling Weight Deflectometer (FWD).

ABS abstracts a ‘core’, typically of 0.9 metres, which is analysed for its chemical and engineering properties. This technique has several advantages over historic trial pitting techniques, one of which is that many samples can be taken in relatively short track possessions.

GPR assesses the thickness and condition of upper trackbed layers and provides information between ABS sampling points. It is a good technique for initial trackbed assessment at a route level but, if used to specify renewals, has to be calibrated using ABS data.

The Springbox, developed by URS, is a 250mm cube filled with test material. This has spring-loaded faces, the forces on which are measured when a pulsed vertical load is applied to the test sample. It is particularly useful for stiffness assessment of ballast for new railways. As will be seen, for rail, FWD is a novel technique still under development.

TRE in action

URS uses TRE for several overseas rail infrastructure asset owners as well as in its national framework contract with Network Rail for trackbed investigation and design. This requires site investigations at over 500 track renewal sites per year with testing of recovered materials undertaken in its dedicated laboratory in Nottingham. The results are then used to design track renewals in accordance with local trackbed conditions.

TRE has also been applied on a route basis, and for track enhancements. Previously, on the West Coast project, TRE was employed to ensure the effectiveness of the ballast cleaning programme. Using desktop studies and high speed GPR calibrated from material testing, a “toolkit” was developed to provide indicative particle size distributions of material layers to assess residual ballast life to optimise ballast cleaning to save millions of tons of material. More recently, TRE was used on the Great Northern Great Eastern upgrade to determine residual ballast life and trackbed limitations. This also tested the suitability of various alternative trackbed treatments.

To enhance operator safety, URS recently developed its Mast Operated Automatic Ballast Sampler (MOABS), which has now received product acceptance following a year of intensive development.

This is used to drive one metre long tubes into the trackbed at depths of up to two metres, and replaces the previous ABS technique. The tubes have plastic liners which hold the sample. A SERB (Specialist Excavation of Railway Ballast) machine collects representative ballast samples for fouling assessment and for Springbox testing.

On the other side of the world, increased iron ore production required a 250% increase in rail traffic with axle loads potentially increased to 40 tonnes on a 426 km long rail line in Western Australia. This required GPR surveys with detailed intrusive investigations at discrete locations to assess trackbed condition and determine the optimum maintenance and renewal strategy including the use of ballast cleaners. TRE has also been applied in Malaysia and Jamaica to ensure trackbed was fit for planned tonnage and linespeed increases.

Measuring the bounce

The idea of testing a pavement by measuring the effect of a weight dropped on it originated in France in 1963. However French (and British) road engineers were sceptical and the following year the idea was taken up by Denmark’s National Road Laboratory which, by 1975, had developed a practical working model of the Falling Weight Deflectometer (FWD). By the 1980s FWDs were in widespread use, particularly in Sweden and the Netherlands. During the 1990s the different interpretations of FWD data resulted in European guidance for incorporation in national standards such as the UK Highways Agency Design Manual for Roads and Bridges.

FWDs provide a load pulse of typically 30 to 120kN via a circular plate in contact with the pavement surface. Deflection measurements are taken at several radial positions and the actual load applied is measured by a load cell. The graph of radial distance against deflection is known as the deflection bowl and provides information on the stiffness of different subgrade layers.

Rail trackbed stiffness tester

Dr Matthew Brough was the right man to present URS’s entry in the innovation competition at the recent RIA conference. As director of pavements, trackbed and materials, he has been involved in trackbed research for the past 15 years. URS was the first company to use an FWD in a rail environment, something that Dr Brough and his team have pioneered. For example, FWD was used in Ireland on over 100 miles of track underlain by peat to assess the impact of increasing line speeds that could have exceeded the trackbed’s critical velocity.

Use of the FWD for trackbed testing requires rails to be unclipped, a beam placed on the sleeper to which the load is applied, and sensors placed on the trackbed. As Dr Brough explained, applying the load to the unclipped sleeper provides detailed trackbed information with no influence of rail and components. In contrast, train-mounted stiffness measurement techniques, although good at identifying problem locations, determine the overall system stiffness making it more difficult to isolate the causes of problems in the trackbed and subgrade.

Thus the FWD provides valuable trackbed data. However, it is currently mounted on a road trailer to which rail wheels are attached for use on the track so on and off tracking is slow and it is difficult to switch lines. As a result, it can only be used during weekend possessions. To make the best use of possession opportunities, the development of a bespoke Rail Trackbed Stiffness Tester (RTST) was proposed.

The intention is to develop an RTST compliant with Network Rail’s product approval within a year. URS considers this requires a three-phase programme: the development of a prototype; further refinement after laboratory trials in a full scale pavement test facility; and field trials to develop industry-ready analysis tools to better understand stiffness, its impact upon performance and remedial solutions.

This would enable RTST to be used in midweek possessions, reducing the cost of FWD testing by 35% and providing an improved understanding of trackbed condition resulting in far greater cost savings from better specified track renewals. Dr Brough also advised that the RTST was the subject of significant international interest with consequent overseas business opportunities. Having made such a clear case for the RTST it was perhaps not surprising that the judges of the RIA/RSSB’s innovation competition decided that it was worthy of a joint first prize.

Ticking innovation boxes

The RTST involves technology transfer from the road industry to directly address a Rail Technical Strategy requirement. It provides improved asset information and potential business for UK plc. It therefore ticks many innovation boxes. As last month’s article on the RIA innovation conference showed, it is one of many real innovations in the rail industry.

It is also one indication of the greater priority now given to trackbed investigations, with the formation of Network Rail’s Track Stiffness Working Group being another example. This group was formed to share best practice, solve problem sites, and define future research in the field of track stiffness.

As Dr Brough says, “If you don’t pay for a site investigation up front, you will end up paying for it anyway,” which is true for any project. Although this lesson now seems to have been learnt for track, there remain possession access constraints. It’s good to hear that, with the development of the RTST, these too will soon be less of an issue.