Catch a train south from Oxford and you’ll see remains of Network Rail’s failed Great Western Electrification Programme (GWEP).
Alongside the tracks are steel piles, some of them driven fully into place, others still sticking from the ground (‘refused’, in engineers’ parlance). According to the plan first announced by the Department for Transport in 2010, these piles should be topped by electrification masts and then overhead lines.
But as the project ran over budget and over time, Oxford’s electric prospects were axed, leaving train operator Great Western Railway using its expensive inter-city bi-mode trains on services that would be better handled as outer-suburban services using its Class 378 Electrostars.
GWEP’s shadow looms large over railway projects. Its problems led to the DfT axing its 2012 plan to wire the Midland Main Line to Sheffield, the trans-Pennine route between York and Manchester, and the ‘Electric Spine’ freight scheme from Southampton to the Midlands.
The first two are once again being talked about with last November’s Integrated Rail Plan, but it means that England has lost a decade.
Meanwhile, Scotland pushed on with its plans and now wants more. Between 2014 and 2019, the country wired 200 miles across its Central Belt, so that all five routes between Edinburgh and Glasgow now run with electric trains. This means that 76% of Scottish rail passenger journeys use electric trains, while 45% of freight in Scotland is electrically hauled.
Progress stalled when the final elements of those 200 miles were completed (Alloa, Dunblane, the Shotts route, and Grangemouth freight terminals in 2018 and 2019). Network Rail’s construction hub at Cadder fell silent and was then converted to a diesel depot for HSTs.
But more recently, work has started to wire the line to East Kilbride and to Barrhead.
More wiring must follow, and Scotland is pursuing a policy of discontinuous electrification using battery-electric trains.
This concentrates initially on Fife. Transport Scotland plans to have Network Rail erect wires from Thornton North Junction outwards on three arms - to reach west to Lochgelly, south to Kinghorn, and north to Ladybank.
There will be a fourth arm when NR reopens the line to Levenmouth.
Two factors drive Transport Scotland. One is decarbonisation and the other is the country’s ageing diesel train fleet. Rail Director Bill Reeve told a Permanent Way Institution (PWI) seminar last April that around 70% of ScotRail’s fleet would become life-expired by 2035. Without more overhead wires, ScotRail would be forced to either keep its diesel fleets running or replace them with new diesel trains.
That includes the Class 156s by 2025, HSTs and Class 158s by 2030, and Class 170s by 2035.
Reeve also listed electric Class 318s and Class 320s (dating from 1985 and 1990 respectively), and even the newer Class 334s, in the 2025 bracket.
Class 156s date from 1987 and work to places such as East Kilbride and the rural routes to Fort William, Mallaig, Oban, and Carlisle via Kilmarnock.
ScotRail’s ‘158s’ (1989) work widely over the network, including commuter trains into Fife and the Borders as well as the Far North Line beyond Inverness.
Class 170s, dating from 1999, are their more modern equivalent. They also work Fife and Borders services as well as inter-city routes, which they share with HSTs built in the 1970s.
Hence Reeve’s conclusion for electrification: “We cannot afford not to do this, and we cannot afford to wait to do this.
"We are struggling to get enough electrification done to be able to replace all of those fleets with pure electrics. We are working to achieve that, but it’s not straightforward, which is why this is urgent.”
Scotland’s Fife plan is modest. From Kinghorn to Ladybank is 16 miles, while running out to Lochgelly adds another seven miles. The section from Haymarket to Dalmeny adds around eight miles. As double track, this equates to around 100 single-track kilometres (stkm, the usual measure for electrification schemes).
NR Scotland Engineering and Asset Management Director Alan Ross also spoke at the PWI seminar. He reckoned that the next six years needed 90stkm of wires erecting, at a cost of up to £2 million per stkm.
He explained: “It’s about driving that cost down. It’s about seeking out those opportunities, reducing the footprint by which it takes us to deliver it, and at the same time working across the industry. Get that core cost down and get that key benefit out, and that will continue to attract, encourage and promote rail as that real critical transport network that can meet those ambitious targets of decarbonisation, but more importantly get that modal transfer from freight in particular to rail. That’s the key prize.”
His call for Scotland to move to a production mindset for electrification merits repetition in England and Wales. So does this argument from him: “The railway is a system of systems. So, the train management and train operation, the infrastructure management and infrastructure operation, and the capital delivery programmes all have to sit together in the same conversation. If they don’t, we’re going to hit bumps in the road. And some of those bumps can be fairly large.”
It was just such a bump that upset Northern’s plans to introduce electric trains in 2018. Problems with the scheme forced NR to delay its delivery. This affected Northern’s driver training plans and led to passengers seeing the benefits of electric trains even later than planned.
Worse still was the political impact of the timetable collapse that followed, which all but destroyed trust in rail companies.
Taken overall, the cost of an electrification programme comes not only in the cost of stringing wires, but also in procuring suitable stock (whether new or cascaded) and training crews. Only when you have wires, stock and crew can services start. That needs co-ordination that is best placed to come from a central mind, likely to be Great British Railways in the future.
But before we get to that, funders (that’s governments) must be convinced that electrification is worthwhile.
Scotland has an advantage here. Its decarbonisation plan for the railway comes from Transport Scotland to Network Rail. Transport Scotland is the Scottish Government’s transport department. In England, all that’s been seen is Network Rail’s advice to the Department for Transport in its 2020 Traction Decarbonisation Network Strategy.
Above all, electrification must become cheaper if it’s to have a chance of receiving government backing and funding.
Three top-level refrains answer the question of how best to make electrification cheaper. The first is a rolling programme. The second is more time, which translates into better access to the railway.
Both have elements of holy grail. Britain has never managed to plan a rolling programme of electrification. It wired its major main lines with decades between projects. The West Coast received wires in the 1960s and early 1970s, reaching Glasgow in 1974. The East Coast followed in the late 1980s and early 1990s. It was then the 2010s before the Great Western followed suit.
Of course, there were other smaller schemes, such as Glasgow’s suburban lines and the London-Norwich route, which built on British Rail’s pioneering AC schemes around Colchester. But never has Britain managed to move sequentially from one project to another in a way that provides the sort of steady work that securely justifies sufficient investment in plant and people to truly drive efficient costs.
That’s not to say Britain should not try. But viewed from 2022, there’s a real risk that instead of a steady flow, Network Rail is about to see Midland and trans-Pennine wiring being delivered at the same time as Scotland’s projects and at the same time as High Speed 2 needs to bring in OLE teams.
What’s absent is the railway’s plan for a rolling programme - one that stacks different schemes against each other in terms of the benefits they release and their costs. Such a plan should take account of wider factors such as rolling stock. This might mean that a weaker line’s case becomes stronger if its stock is approaching the end of its life and needs replacing.
The Railway Industry Association has been working on this for northern England, and expects to publish its findings around the time that printers produce this issue of RailReview. That’s surely a good start, but it needs Great British Railways to take this forward across the rest of England.
The second refrain is the perpetual call for more time. It’s entirely logical that work is more efficient if there’s no need to keep stopping and starting, so longer blockades are better than night or weekend closures.
Unless you’re a passenger wishing to travel. As Alan Ross told the PWI: “The railway is here to move people and goods. The railway is not here to dig the railway up and create new projects.”
So, there will always be tension between running a service to earn money and closing a route to spend money more efficiently.
In some cases, NR will need to invest in alternative routes to deliver reliable links between towns and cities. It’s recently done that on the Calder Valley line that joins Lancashire and Yorkshire via Todmorden. This gives train operators, particularly TransPennine Express, an alternative to its usual route via Huddersfield.
Taken to its extreme, HS2 has the best chance of delivering efficient electrification. It will be working on greenfield sites with no trains to worry about.
At the other end of the spectrum is East West Rail, which was planned to be electrified, but then descoped to be a diesel railway. When even miserly ministers realise they can no longer resist electrification (for all those well-known decarbonisation reasons), it will cost them more than had they done it while the line was being restored.
The third refrain can best be described as ‘measure twice, cut once’. In other words, plan the work thoroughly before starting construction.
This is perhaps the biggest lesson to learn from Great Western’s electrification, where pressure to deliver on time led to work starting before plans were properly developed.
This showed itself most clearly in the problems the project had with piles. The National Audit Office (NAO) summarised the problem in a report: “Network Rail did not carry out sufficiently detailed surveys of the route before the ‘detailed design’ took place. This is critical, since if ground conditions at one site are not as expected, designs for a number of nearby locations could need to be changed. This delays piling and installation of masts. In November 2015, Network Rail estimated that 78% of designs completed so far had needed to be revised.”
Scotland is now using test probes at each pile site to check whether the ground is suitable. This might appear to be extra work, but it costs less than the disruption to a schedule that comes with piles refusing to drive.
Being sure in advance that every site selected for a pile is suitable helps drive the production mentality that Alan Ross espouses. It was this mentality with which NR started its Great Western programme.
It procured a special train, called HOPS (high-output piling system) to do just that. It could drive up to 30 piles per night. Behind it, NR could send another train to lift masts into place and then fit booms. These booms hooked into place, making them quick to fit. The crane didn’t need to hold the boom in place as teams fitted bolts - it could move straight to the next mast.
But this production line fell foul of piles refusing to drive into the ground. It meant that project managers had to revise plans to cope, shedding their production line in favour of bespoke work to particular piles and masts. With this, any hope of keeping to their schedule disappeared.
So, NR had the right idea, but delivered it badly. It should revisit HOPS for future projects, but only if it does the proper groundwork first. (Some of the HOPS kit remains dumped in Swindon.)
Despite having a clear view that the Great Western modernisation was introducing electric trains, the DfT did not do enough to link both sides of the railway system. The NAO was very critical in its 2016 report: “ When the Department entered into a contract to buy the Intercity Express trains in July 2012, creating fixed deadlines for electrification, Network Rail had only just identified that it would need to develop a new type of electrification equipment. It is unlikely that either Network Rail or the Department had a good enough understanding of the work involved in developing and installing this new design, to be confident in the time it would take when the Department let the contract for the new trains.”
The NAO explained that the DfT was at one point facing bills of £400,000 per day for Hitachi’s new trains, that couldn’t run until NR completed its wiring work. DfT also had to negotiate a variation order to its deal with Agility Trains to convert the 21 electric trains it had ordered into bi-modes, to run on diesel or electricity.
Proper management from the outset of the project as a combination of wiring and trains would avoid these extra costs.
There’s another aspect of proper management that NR must consider. That’s deciding what it wants from electrification.
The flip side of this is being clear about what isn’t electrification. Renewing signalling isn’t electrification, yet its costs can too easily be lumped into the bigger bill, which harms electrification’s overall case. Likewise, raising bridges to give increased freight gauge. It’s worth doing and worth spending money on, but it’s not electrification.
Across recent schemes, it’s generally reckoned that foundations, masts and wires consume about one-third of project costs. Another third goes to civil engineering work to stations, bridges, track lowering and tunnels. The final third goes to signalling immunisation, traction power connections and compensation to operators.
Bridge clearances remain a contentious problem for electrification schemes. With care and engineering judgement, it’s possible to cut the costs of electrification by cutting the number of bridges that need rebuilding. This care and judgement must come from the client. Leave the decision to the contractors and they will likely recommend more work to earn more money.
So, compare the 270mm static clearance mandated on the Great Western project with the 85mm clearances that exist now (and have for many years) on the Great Eastern Main Line out of London Liverpool Street.
If real life tolerates 85mm, then perhaps your bridge (for which the best you can get is currently, say, 170mm static clearance), doesn’t need rebuilding.
Might you consider other options? You might look at some of the recent developments in probabilistic gauging, which provides a better estimation of clearances in the real world than you find by simply adding maximum figures for each part of track and train as they move in relation to a structure such as a bridge.
Engineers at Cardiff Intersection Bridge couldn’t easily rebuild it to give the mandated clearance. Here, the main line crossed a river while being crossed itself by another rail line. The answer was a combination of surge arrestors and insulation that NR claims saved £40m.
Several engineers told RailReview that rebuilding bridges cost millions, while surge arrestors came in at tens of thousands. That’s enough to prompt NR to look very closely at what it wants as a client.
For NR’s trans-Pennine and Midland Main Line wiring projects, this switch to surge arrestors (in a technique properly called voltage-controlled clearances, VCC) has removed 80 bridge renewals from those projects’ scope. This includes the challenge of Stalybridge Tunnel, where NR was at one stage faced with a six-month closure.
It’s not just clearances under the bridges that add cost. There’s a requirement to make parapets on bridges crossing over electric railways 1.8 metres high. But simply adding height to parapets can trigger extensive reconstruction if the bridge is not sufficiently strong to cope with the extra load (either the weight or the forces that come from increased wind loading).
Engineers argue that these higher parapets are not proportional to the risks involved (chiefly, electric shocks to pedestrians). Some bridges in some areas will need higher parapets, but (as is so often heard) one size does not fit all.
Both clearance problems are well recognised within NR. Following considerable work, it now suggests bridge soffit heights of 4.25 metres above the rail, rather than the nearly five-metre heights seen in typical Control Period 5 projects.
Thus is NR beginning to shed its reputation as a poor client. This reflects Chief Executive Andrew Haines’ comments to RIA’s annual conference in October 2020: “I see people in Network Rail who are gold-plating schemes in the first instance because our standards have not been driven to reflect a cost-constrained environment.
“I see designers who look to off-load risk from designers to the customer effectively by over-engineering, and that’s a large part of the story for Great Western electrification, and the fact we were a poor client for that doesn’t mitigate the fact of that poor behaviour.”
It’s not just bridge clearances that challenge. The distance from a train’s pantograph to passengers on platforms also matters.
Here, nearby bridges may well push the wires unacceptably low through platforms. One option could be to insulate the outer horns of pantographs where they come closest to passengers. For a railway split between track and train, this answer needs co-ordination between both sides.
Life becomes a bit harder for NR when it comes to deciding what’s wanted from electrification. What trains? What timetable? What power draw?
It’s a question better asked of rail’s controlling mind - supposedly the DfT for the last two decades, but soon to be Great British Railways. But it’s not easy to answer. It depends on forecasts of likely traffic. It depends how efficient the funder wants spending to be. Does that mean installing just enough power for the current service? Or sufficient to provide room for expansion?
Glance back at GNER’s timetable of services from Edinburgh from 1997. There were 21 weekday trains in mid-summer. A handful were diesel HSTs, but most were Class 91 electric locomotives hauling Mk 4 stock, as British Rail introduced when it electrified the route earlier that decade.
Today, GNER’s successor, LNER, runs 28 trains per day south from Edinburgh. All have the ability to run electrically. Add Lumo’s five daily trains (which must run electrically), and TransPennine Express’s four trains (which could and should, but don’t run on electric) and it’s easy to see why NR is busy improving power supplies between Edinburgh and Newcastle.
As NR starts work on wiring the trans-Pennine route between York and Manchester, the question of power is pertinent.
TPE already has a fleet of bi-mode trains, similar to those used by LNER and GWR, so their characteristics are known. But it also runs diesel services along the route, as does Northern. The future for these is less known. As is the future of freight services across the route. NR’s planning provides for more freight paths, but it’s not clear whether they will be diesel or electric.
There should be savings in the way NR brings power to the railway. Rather than classic grid connections, there’s scope to use static frequency converters (SFCs) to feed electricity into the overhead lines. Using SFCs reduced the likely cost of the second phase of the ECML’s power upgrade from £1.3 billion to £400m.
Once there, the power could be fed along OLE that carries higher mechanical tension, which in turn allows masts to be spaced further apart. This cuts the number of masts and foundations needed - albeit the masts and wires might need to be beefier to cope with the higher tension.
Siemens suggests that its SICAT range could stretch mast distance from today’s 65 metres to up to 80 metres, and perhaps even 110 metres with its SX range, although it’s not approved for UK use.
Perhaps this is the sort of innovation that Andrew Haines meant when he addressed a Permanent Way Institution meeting last year: “Critical to future electrification will be a relentless wave of innovation and collaboration to make it more affordable. That demands a different approach from the one Network Rail has relied upon in the past. It demands a different sort of contractual relationship and a different sort of appetite from the supply chain to offer up innovation.”
While Bill Reeve in Scotland senses urgency in replacing ageing diesel trains with new electric stock, in England there is less pressure.
English operators such as Northern face the same challenges. It still runs the same ‘156s’ and ‘158s’ that Reeve’s ScotRail operates. Northern’s are just as old, but there’s no talk of wiring Newcastle to Carlisle for its roughly three trains each way every hour, for example.
Yet Haines is clear about the importance of electrification within decarbonisation. He told that PWI meeting last year: “Decarbonisation is one of our biggest opportunities and one of our biggest challenges. We get our approach right and we secure the future of our country’s railways as the UK’s greenest form of mass transit for generations to come. If we get it wrong, we lose ground to our competitors and throw away one of the railway’s major selling points - its sustainability.
“I was pleased to see the government support and (I quote) ‘ambitious, sustainable and cost-effective programme of electrification’ in its transport decarbonisation plan. But it would be a massive mistake to think we now have a blank cheque to decarbonise the remaining 60% of our railway.
“The ball is now firmly in our court to demonstrate to government that the railway is capable of delivering value for money for every pound spent on decarbonisation.
“Demonstrate that we have learnt from the lessons of the past. It would be stupid to underestimate the damage that has been done by the cost and schedule overruns on the Great Western electrification, to name one project. Demonstrate that we can be trusted to deliver on schemes efficiently and at reduced cost.”
Measure twice, cut once.
