Doncaster punches above its weight. From some angles, it looks like a dozen other Yorkshire towns – Rotherham, Castleford, Pontefract – but its position is significantly elevated by its strategic importance as a railway hub. It’s Crewe, but the right side of the Pennines.

With six routes converging, Doncaster’s population can fluctuate by several hundred every few minutes as a plethora of operators make transient appearances, not least on the East Coast main line. And this places a heavy burden on project and engineering teams: getting it wrong here will cause far-reaching disruption, courtesy of the ripple effect.

Time and place

Fulfilled by the S&C North Alliance – a collaboration between Network Rail and AmeySersa – the town’s Christmas track renewals were extensive and logistically complex, spanning two sites. One, immediately south of Doncaster station’s east-side platforms, encompassed crossovers between the Up Fast and Up East Slow, as well as three turnouts into Platforms 1 and 2 and carriage sidings. The other, half-a-mile further north at Marshgate Junction, involved three more turnouts and a crossover with the Up Fast, Up Slow and Thorne lines being affected. Both were like-for-like replacements of existing layouts, mostly dating from the late 1970s.

Much of the work was programmed for an all-lines- blocked period with accompanying OLE isolation, booked from 23:00 on Thursday 24 December until 09:00 on Sunday 27th. Thereafter, the Up and Down Fast and Up and Down Doncaster would be handed back, together with lines serving the west-side platforms, allowing the resumption of services to/from York and Leeds using Platforms 4-8. The Up Fast was blocked again for 8 hours overnight Sunday- into-Monday and possession retaken, with the Down Fast, through the early hours of Tuesday 29th. All lines were due to be given back at 06:00 on Tuesday. This track access regime shaped the phasing of the works.

What ifs

We all know – at least we should – that key to the successful delivery of any such project is thorough planning; it’s also widely recognised that, historically, the railway has not been particularly good at it. Things have improved hugely since the overruns of Christmas 2007 with the introduction of better processes, although critics seem reluctant to acknowledge this.

Of course, the cause was not helped by the scenes – and subsequent spleen-venting – that resulted from the failures at King’s Cross and Holloway Junction during Christmas track renewals a year ago. This, you will recall, resulted in several hundred passengers queuing on the road outside Finsbury Park Station on 27 December 2014, in some cases for three hours. They then had to contend with seriously overcrowded trains. It was an unfortunate episode and inflicted reputational damage on Network Rail. Broadcasters still wheel-out the pictures periodically as a stick with which to beat it.

Primarily, the problems there arose from plant/ equipment failures associated with the removal of scrap materials, the resulting misalignment of the physical works with the logistics plan, and the impact this had on train crew availability. As you’d expect, Doncaster shared many of the same challenges, sometimes on a larger scale: 7,500 tonnes of ballast to replace, 6,000 tonnes at Holloway; 23 engineering trains to manage, 14 at Holloway. But lessons have been learned in the past 12 months, bringing more robust contingency arrangements and mitigation measures. Network Rail standard NR/L3/ INI/CP0064 (Delivering Work Within Possessions) provides a framework for these, stipulating, for example, the duration of floats (periods of unallocated time) that must be built into the programme so delays can be clawed back.

Step by step

Across the two Doncaster sites, the work encompassed an effective renewal length of 1,535 metres, with 52 plain-line panels – arriving on Salmons – nine spine panels and 35 S&C panels. Of the latter, 18 were brought to site on tilting wagons, the remainder being pre-delivered by HGV and positioned optimally as part of the preparatory works. They were all manufactured, in modular form, by Vossloh Cogifer in Scunthorpe.

At Marshgate, progress was made in three distinct phases, driven by the staged handbacks. Phase 1 incorporated turnouts and a linking 120-metre plain-line section on the Up Slow, but in sharpest focus was 2481 points on the Up Fast (Phase 2) which was planned to be tamped, commissioned and clamped by 04:15 on Sunday morning. This included six hours wheels-free for signal testing by TICS Global and left 4 hours 45 minutes as a float and for handback. From 09:00, the line would reopen with a 50mph TSR imposed. In the event of significant delays, a cut-and-run option was available whereby the points would have been plain-lined using six standby panels to allow implementation of the planned Christmas timetable.
Fettling [online]

Being a 70mph turnout, the G-switch panel for 2481 points exceeded 30 metres in length and delivery required an abnormal loads movement order, as well as the temporary widening of a site access gate and provision of a 500-tonne mobile crane to offload and position the panel in Marshgate Sidings. Installing it involved meticulous planning between the project team and the VolkerRail Kirow team to undertake the required lifts using two KRC250 Kirow cranes in tandem lift mode which necessitated bridge loading checks on UB332 over the river Don.

Completion of the crossover between the Up Slow and Down Thorne (Phase 3) took place overnight Sunday-into-Monday as it required a spoil train to be loaded on the Up Slow with the adjacent Up Fast blocked, the site being protected under Any Line Open arrangements.

Another phased methodology was established for the station site. Here, all of the old track was scraped out and reballasting completed before any relaying took place, beginning with the platform lines and working outwards. The panel installation started on the Up Fast (Phase 1), then the Up East Slow (Phase 2) to finish the crossovers, and finally the platform/carriage siding turnouts (Phase 3). Again, this involved the Kirows working in tandem due to the typically 16-tonne panel weight and 12-tonne modular lifting beam.

Generally, the formation treatment was for 300mm new ballast depth with a geotextile PW4 separator and geogrid, but the contingency plan allowed for this to be reduced to 200mm on the Slow/platform lines or skim as a last resort.

Breathing space

In terms of plant and manpower, the two worksites were largely independent, with separate train plans able to accommodate a deviation from schedule of between two hours ahead and four hours behind. An emergency light engine was located at each site, a number of trains topped- and-tailed with drivers both ends, and mostly local crews sourced. Further redundancy was built-in to account for machine breakdowns, with a spare dozer and RRVs available.

The two Kirows were shared, first completing Phases 1 and 2 at Marshgate before transiting to the station to relay Phases 1 and 2 there.

An 81⁄2-hour float was built-in between these activities. One crane then returned to Marshgate for Phase 3 (the float being five hours) whilst the other completed the station works. Two VolkerRail Matisa B41 Tamping Machines serviced both sites to a similar pattern, completing in excess of 3,200 metres and 16 switches of tamping.

To guard against failure of a Kirow crane, which would have severely impacted the overall plan, the project team arranged for two additional Kirows – being used on the S&C renewal at Haymarket, west of Edinburgh Waverley – to be routed to site for contingency purposes, ready to start lifting from midnight on the 27th. The interim handback had been assessed to ensure the Christmas timetable could operate even in this event.

One key advantage Doncaster had over Holloway was the track layout, offering more routing options and thus reducing any overrun impact. Beyond this, no other major works were ongoing within the possession, so there was no need to stack trains on site. Nevertheless, the criticality of this project was not under-estimated and a 95% probability for on-time handback had to be achieved based on a Quantitative Schedule Risk Analysis; this was greater than the 90% demanded by the Standard. Network Rail also recognised the need for a greater level of senior managerial support on site than had previously been allocated.

So, that’s the theory outlined. What about the practice?

Boots and shovels

With the possession taken on time at 23:00 on Saturday, excavation and reballasting at the station ran broadly in sync with the planned timeline. However, a late Form C and machine problem at Marshgate caused a delay of around two hours, resulting in a contingency option being invoked at one of the hold points; this reduced the Phase 1A (2476B, 2465B, 2480 points) dig from 300mm to 200mm, saving two hours. Phase 1B also fell behind due largely to access constraints whilst panels were being laid behind.

Work on the critical 2481 points got underway 13⁄4 hours late. This had become 21⁄2 hours by the time all the panels had been installed, but that was pulled back when the top-stone was dropped and during tamping. However, the signal testing took longer than expected, a result of welder trolley movements dropping track circuits and the running-out of a weld on the River Don underbridge which required a closure rail to be installed. “That’s what the float’s for,” asserted Steve Varley, Network Rail’s senior responsible engineer (track). It did, however, cause some nervousness as the signalling was not signed back in until 08:00, but still with an hour in hand.

The Kirows then moved to the station site and the relay there got underway. Again, things essentially followed the expected timeline from start to finish during daylight on Sunday. Back at Marshgate, the Down Thorne was uplifted early but excavation and reballasting had to wait until overnight possession was retaken of the Up Fast; this was delayed 40 minutes by late-running flood-affected trains. Relaying continued following handback at 06:00 on Monday, with wheels-free testing being carried out in stages across the two sites, including confirmation of detection on 2481 points after unclamping. This process again suffered delays but final handback was ahead of the booked time at 06:00 on Tuesday.

What the Dickens

For those involved at the trackface, this could be described as just another job. It went more-or-less to plan, notwithstanding the occasional – and inevitable – hiccup. That’s the real world for you. Much was achieved during the floats, flooding the site with welders to complete 222 welds, rather than relying on temporary clamps as would normally be the case. This enabled the 50mph TSRs to be removed after only one week, rather than four; seven stresses were also completed.

Steve Varley made the point that the good relationship nurtured by the Alliance with Network Rail Ops, the TOCs and FOCs, paid considerable dividends. Their understanding of the plan – and buy-in – ensured help was given with the likes of late train policies to ensure possessions were granted on time. And the signallers in Doncaster PSB are also singled out for praise, assisting with route setting and line blockages when needed. This all contributed to the project’s success story.

It’s not unreasonable for the media to highlight the railway’s occasional shortcomings when it comes to delivering its engineering works. They can negatively impact on people’s lives and it should be held accountable for that. But the industry can and does routinely get it right; Doncaster is testament to that. Trouble is, good news is no news. We can, though, hope that the Ghost of Christmas Past has now been put to rest.