In the complex story of the electrification of the Great Western main line, one name keeps cropping up – Steventon Bridge (issue 167, September 2018). This structure, and the problems it has caused, has been the subject of much media coverage, both technical and public, and has led to some robust discussions with neighbours and stakeholders. It has also driven an important engineering challenge.
Rail Engineer was invited to the Railway Technical Centre at Derby to meet Garry Keenor of Atkins, to discuss the design and construction issues at the site and hear how the challenges were overcome. Garry described the process as a new approach to modelling the OLE/pantograph interface using FEA (finite element analysis) techniques.
It sounded fascinating.
The Great Western Route has undergone a massive change in recent years as part of the Greater West upgrade programme. Following on from the successful introduction of diesel-powered High Speed Trains in the 1970s, this latest project sought to improve journey times and the passenger experience still further by electrifying the route from London to Cardiff using a 25kV AC overhead contact system (traditionally known as overhead line equipment – OLE). The scheme has faced many challenges through the years – one of which being what to do about Steventon bridge.
The village of Steventon is situated to the West of Didcot on a 125mph linespeed twin-track section of the Great Western main line. The village is cut in two by the railway, and the line is traversed by three roads that cross the railway using Steventon Bridge, Stocks Lane level crossing and Causeway level crossing. Steventon Bridge is a Grade two listed brick-built arch structure dating back to the era of Isambard Kingdom Brunel – a valued piece of history but subject to much remedial work over the years.
Safe and reliable OLE operation depends on the ability of the pantograph on the train to collect traction current and follow the wire as it rises and falls to meet infrastructure conditions. Particularly relevant to Steventon are the restrictions imposed by the two level-crossings and the low bridge height.
UK standards generally comply with the EU technical standards for interoperability but, for OLE, a generic set of OLE rules apply, with allowed non-compliance for UK conditions. To achieve the best possible performance under all conditions and at all locations, without the need to design from first principles all the time, these rules must necessarily be conservative.
Under these rules, the designer is driven to include the rate of rise and fall of the wire as seen by the train, generally expressed as a gradient of 1:X (one unit of vertical rise and fall over X units of length). Deviating from limits can mean loss of collector equipment contact, and therefore traction power, along with increased wear on the contact wire, arcing, and a higher potential for dewirement.
The early design processes made the normal assumption that the bridge would be reconstructed to allow for a ‘standard’ OLE design and the running of electric trains at normal linespeeds, in this case 125mph.
However, as the processes of consultation with local authorities and stakeholders progressed, it became apparent that this assumed strategy of reconstruction would not be accepted and the existing bridge would have to be retained, resulting in the linespeed for electric trains being considerably reduced and journey times being lengthened by around five minutes.
As the existing Steventon Bridge is very low, the installation would result in a wire height of 4.22 metres, even with the adoption of reduced electrical clearances. That would not normally be a problem, but Stocks Lane level crossing is only 399 metres away, where well-established safety legislation requires the provision of a minimum contact wire height of 5.8 metres, resulting in a wire height of six metres at the adjacent structures.
To meet these two constraints, the wire gradient between them could be set no shallower than 1:202. The normal UK-applied rules are that the gradient is no steeper than 1:(5 x linespeed in mph), so a 125mph linespeed would dictate a maximum gradient of 1:625. (It is interesting to note that mainland European systems widely use a maximum gradient of 1:1000.)
Working backwards, the 1:202 gradient would impose a maximum linespeed of 202/5 or 40.4mph – a speed which would impose crippling restrictions on this key trunk route to the West, taking into account braking and acceleration. Because of this, the engineers involved used empirical judgement and set the limit at 60mph.
The programme identified this problem as early as 2010, and subsequently sought to reconstruct the bridge with a higher profile – as has been carried out at many other sites on the route. However, permission to demolish and reconstruct was refused by the District Council.
By 2018, Network Rail was facing imminent entry into service of the Didcot to Swindon section of the electrified railway, and so had no choice but to construct the OLE at Steventon with the gradients set at 1:202. Following this, Great Western Railway made the understandable decision to switch its bi-mode trains from electric to diesel and back again either side of the bridge, and slow its non-bi-mode, electric-only EMU fleet down to 60mph to pass under the bridge.
The result was significant time penalties – especially in the case of westbound trains which stopped at Didcot and then had to accelerate to linespeed on diesel power. The situation was acceptable at first, as trains were operating to the historical diesel timings, but the approaching December 2019 timetable change meant that a solution was needed to meet the planned overall London to Cardiff timings.
Incidentally, the designers were not simply taking a leap of faith in setting the speed limit at 60mph. They knew that the traditional generic rules on gradient were conservative, using values that could deliver minimum and maximum contact forces at the pantograph and comply with national rules. However, that philosophy does not mean that fiercer gradients will necessarily drive excessive forces. Network Rail’s new Series 1 OLE, developed specifically for the Great Western route although intended to be used elsewhere as well, uses higher tensions than normal and also had been designed in parallel with the new pantographs on GWR’s electric fleet. Atkins was responsible for the detailed design of Series 1 at Steventon and both Network Rail and Atkins believed that higher speeds were possible with the existing gradients, but needed to demonstrate that and achieve a satisfactory case for those higher speeds.
D-RSS simulation package
Atkins had recently developed the Dynamic Rail Systems Simulation (D-RSS) software package and it was agreed with Network Rail that it could be applied to evaluate and test the potential for operation at a higher speed at Steventon.
D-RSS simulates the pantograph/OLE interface at a higher level of fidelity than traditional lumped-mass model systems, using a modern FEA approach. It was developed in-house using FEA software for holistic dynamic simulation across the system and has been validated to BS EN 50318:2002.
The software offers optimised design development, useful for areas where bespoke design arrangements are necessary, and challenges conservatism in system design, resulting in a leaner, optimised and location-specific design solution. A sister tool, OLE-StAT, gives the opportunity, in the later stages of the project and during commissioning, to compare the as-fitted system with the original design.
Overall, D-RSS provides the opportunity to determine where the TSI compliance threshold really lies, rather than just following rules which may be too conservative. In the case of Steventon, the modelling predicted that the compliance limit would be reached at around 110mph.
The results of the Class 800 simulations using D-RSS gave Network Rail the confidence to progress to the next stage, high-speed testing. So, in early 2019, a twin five-car Class 800 set, with two pantographs raised and spaced at 214 metres, passed through Steventon at gradually increasing speeds in both directions, starting at the existing temporary linespeed of 60mph.
The pantographs were instrumented using DB Systemtechnik equipment by DB ESG employees. Atkins staff on the train then analysed the outputs immediately after each run, using pre-prepared routines. The results were shared in real time with ground-based staff to allow a go/no-go decision to be made immediately before the next run.
Almost immediately, the results were found to correlate closely with the D-RSS predications and a clear relationship between speed and contact force was identified. Run speeds were slowly increased until, finally, two passes were undertaken at the full 125mph design linespeed. However, at these speeds, the contact force was starting to exceed the allowable limit in the TSI and national standards, so it was decided that 110mph was the acceptable limit.
The second type of electric rolling stock introduced onto the Great Western route is the Class 387. This time, the test train consisted of three five-car sets with a total of three pantographs, 80 metres apart. For this class, the speeds were raised from 60 mph to 110mph and, again, the test results correlated closely with the D-RSS simulation.
These runs confirmed that the higher tensions of Series 1 OLE, combined with the compliant nature of modern pantographs, resulted in a performance well beyond those limits imposed by the current UK rules.
As a result, Network Rail felt able to raise the permanent speed restriction for electric trains to 110mph in September 2019, thus allowing the speed capability to be available for the December timetable change. In addition, Class 800 trains no longer have to switch to diesel power and back again.
However, while this is a very positive result, the investigation is not over. Network Rail suspects that it might see increased contact wear, and therefore higher maintenance costs, due to the higher forces at the site. Network Rail will therefore undertake detailed wear measurements throughout the graded section every six months, the outcome of which will allow re-assessment of the adopted periodicity of wear measurement or, indeed, may allow further increase in the allowable speed.
So, Network Rail and Atkins have been able to use the D-RSS tool to raise electric train speeds from a severely restricting 60mph to a more palatable (but still reduced for class 800) 110mph. This is a positive achievement, but still does not meet GWR’s requirement to utilise the full capability of the route and run at 125mph – only reconstruction can achieve that.
However, raising the linespeed to 110mph, without further modification to the infrastructure, is a great achievement.
In overviewing the work, senior Network Rail project engineer Simon Warren said: “The D-RSS system moves dynamic OLE simulation from a process that is very challenging, and therefore rarely done, to one that can be a routine part of a design process in constrained locations, removing significant capital cost and programme time from major programmes of work.
“Following the example here, D-RSS may be used to analyse a location where it is not possible to comply with wire gradient rules between a level crossing and a low overbridge, thus limiting electric train speed. D-RSS can be used to support the lifting of these speed restrictions, by demonstrating that electric traction performance will be compliant – even though the gradient rules may not be met.
“This methodology can be used in the future at other locations as a tool to avoid other constraints, thereby reducing the capital cost of electrification.”
Thanks to Garry Keenor, group electrification engineer at Atkins, independent consultant Peter Dearman and Simon Warren of Network Rail for their help in preparing this article. The work of Doctor Nikolaos Baimpas, now of Train-Rail Infrastructure Solutions (T-RIS), in developing the D-RSS system and applying it to the Steventon problem is gratefully acknowledged.