WILLIAM BARTER suggests the Revised Business Case for HS2 shows some muddled thinking
HS2: visual of how Birmingham Curzon Street might look
. Courtesy HS2
Notice to Proceed for the main works contracts for HS2 is very welcome news, and the Business Case that supports it gives a much clearer impression of how HS2 will operate than its predecessors.
So far as HS2 itself is concerned, the business case outlines the Phase 1/2a ‘Statement of Intent’ service. This is a careful bit of wording recognising that powers currently exist only for Phase One infrastructure, with all trains to the West Coast main line (WCML) operating by way of the Handsacre junction (the link joining HS2 to the Stafford line); however, the Phase 2a infrastructure is intended to be brought into use simultaneously with Phase One.
One change since previous Business Cases is an hourly HS2 service between Euston and Stoke-on-Trent, serving Stafford and extending to Macclesfield. This is good not just for Stoke: this train picks up the Stafford call and both Liverpool trains in the hour can take the faster route via Crewe. One of these is to detach at Crewe a portion for Preston, which it is also suggested could be extended to Lancaster.
Another change, arising in the Phase Two service, is that instead of Euston – Scotland trains splitting at Carstairs for Glasgow and Edinburgh, the splitting location has shifted to Carlisle. Logically this would be a passenger stop as well, and if it is to be a passenger stop in Phase 2b, equally logically the Phase 1/2a Euston – Glasgow service calls there. So, together with the Lancaster train, two significant locations that risked losing their direct trains to London in previous specifications will now have them provided by HS2.
‘One of these and one of those’ strategy: ‘Statement of Intent’ interim service on HS2 for 2029-31.
Courtesy HS2
OLD OAK
It is now clear that Old Oak Common is expected to act as the London terminus until Euston is ready. Thankfully this may be only a short period – the most optimistic timescales suggest only just over a year, but it would be unwise to bank on that. What sort of service could run during that period? The Business Case actually implies a classic service to Euston and an HS2 service with all passengers transferring at Old Oak are equivalent in terms of demand and benefits, so to be worthwhile the services should offer some benefits over and above the HS2 trains themselves, and avoid leaving gaps on the classic network. Multiple consecutive timetable changes on such a complex route as the WCML must also be avoided.
Three trains per hour between Old Oak and Curzon Street? Fine, that works as they are confined to HS2 so can be timed at will and would be an enhancement to the present service. Do they, however, form an adequate replacement for existing trains to Euston, to the extent that those services can be reconfigured around the needs of intermediate stations? Probably not, and in those first couple of years it would be wise to have the capacity of the WCML to fall back on in case of some ‘Britannia’-like problem requiring suspension of HS2 services (the first 25 of the newly-introduced ‘Britannia’ steam locos were temporarily withdrawn soon after introduction in 1951 due to reports that the driving wheels had shifted on their axles). But maybe the headline ‘flyer’ – the 07.30 from New Street – becomes superfluous, if we reasonably assume it carries the briefcase brigade who will be happiest to change to Crossrail for the City and Docklands. It should be simple enough to put the standard stops into the 07.30 and thus fill a key gap at present for a service from the West Midlands to suit commuters to Milton Keynes. Moreover, with the high value business now on HS2, the peak ‘set down only’ and ‘pick up only’ stops at Milton Keynes and Watford by Birmingham trains could be opened to London commuters, directly adding London commuting capacity for those places at the earliest opportunity.
It becomes more awkward to choose any other services to terminate at Old Oak. The Business Case suggests a total of six might terminate – operationally this is fine, as it boils down to one train per platform per hour, which is prima facie feasible. It would be best if they could be timed into their ultimate paths, so the trains could simply be extended to Euston in due course without a further timetable change on the classic railway; turnrounds of 45 minutes or so at Old Oak would allow that.
Pick ’n’ mix: off-peak service proposed by HS2 once the western leg of the high-speed line is open. Courtesy HS2
But which trains? A scenario outlined in the Business Case would see one Liverpool (out of the two per hour planned by Avanti West Coast) and one Manchester service terminating at Old Oak, plus the hourly Glasgow. But this looks like a classic case of planning by playing trains on paper. They must be existing services diverted, rather than additional trains, as capacity to run extra trains beyond HS2 does not exist. Picking ‘one of these’ and ‘one of those’ simply creates inefficient use of platforms at Manchester and Liverpool by disrupting the rolling stock circuits, and leads to uncertainty as to which station to use – do you head for Old Oak, just miss one and head back to Euston only to miss the next there, or do you wait an hour at Old Oak?
Far better would be to convert one service group to HS2 in full. That could not be the Manchesters, as doing so would deprive intermediate stations of services, but with the Handsacre link in place the two Liverpools per hour could be diverted to Old Oak and still serve Stafford. Something would need to run in the classic path south of Crewe in the peaks to maintain fast peak services for Trent Valley stations currently provided by a Liverpool dropping in, but that should not be a major timetable upheaval. The Glasgow might also work, if you really think the passengers most characterised by heavy luggage, and for whom the journey time saving is proportionately least, will be happy to struggle onto Crossrail instead of taking taxis from Euston.
The same ‘one of these’ and ‘one of those’ approach is apparent in the scenarios presented for classic services (overleaf). For instance, Watford is shown to have a stop by a Manchester train in addition to its present hourly fast Birmingham. But Birmingham by virtue of both size and proximity must be the prime generator of traffic – would it not be better to have a half-hourly service to Birmingham and make the Manchester link with a decent connection at Milton Keynes, which would still be an improvement on now?
Pendolino on song: No 390156 works the 17.40 Euston to Manchester service at Grendon on 29 May 2020. Jamie Squibbs
OFF-PEAK
The off-peak service proposed in the Revised Business Case manages to be both operationally complex and commercially unambitious. Good points are that Milton Keynes has six fast trains per hour to Euston – which should be spaced at regular intervals, instead of all of them in half the hour as occurs now – and a half-hourly fast service to the West Midlands. But the Trent Valley does not seem to gain anything, whereas an area that suffers from HS2 construction whilst not being served by HS2 should expect to gain from released capacity. Nor does Northampton’s service appear better than now, which is odd as the Business Case quotes Northampton Borough as supporting HS2 on the grounds that its service will improve!
The difficult train to path is the one that runs via Northampton but apparently has to make Fast line calls at Watford, Hemel Hempstead and Berkhamsted. Pathing this one at all means either bunching the fast trains, detracting from what should be a regular interval to Milton Keynes, unless you spray pathing time at them. Meanwhile the capacity consumed by the stops stands in the way of improvements for the Trent Valley and Northampton.
PEAK
If the off-peak service is pick ’n’ mix planning in action, the peak service is inspired by Jackson Pollock! Station stops are thrown seemingly at random onto the diagram, hoping some will stick. Five trains per hour to Tring have three different stopping patterns. Six trains per hour to Northampton do four different things on the way. Watford loses the call by a fast Birmingham train that it has now. Milton Keynes, having gained a half-hourly fast service off-peak to the West Midlands, reverts to hourly in the peak.
If the commercial rationale is obscure, the plan does not seem to be operationally rational either. On the Fast line, there is no pathing reason why the same stops as off-peak should not be made at Watford and Milton Keynes. Elsewhere, Von Moltke’s law applies! No plan of battle survives first contact with the enemy; no train service specification survives first contact with the graph paper. Not one but three trains per hour are now trying to make stops on the Fast line, and that just doesn’t fit. Even by flighting the other trains and thus destroying the clockface service, one can be run as far as Ledburn on the Fast before crossing to the Slow line, one to Bourne End, and the third not at all. With all these stops, the number of Fast line departures from Euston is just 12, whereas 14 should be feasible.
On the Slow line, if five trains per hour are to terminate at Tring, they must be at pretty much fixed 12-minute intervals on arrival there, just enough to reverse and clear the platform before the next turns up. So that sets a structure for the whole service; owing to different stopping patterns, it makes the service irregular at Euston. Whilst a semi-fast Bletchley or Milton Keynes train can fit in the gap between two Trings, trying to fit four trains into five gaps makes half-hourly intervals impossible, and the irregular gaps between Slow line trains makes it far from certain a Fast to Slow move at Ledburn or Bourne End will find a Slow line path when it needs one.
Jackson Pollock: evening peak service proposed by HS2 once the western leg of the high-speed line is open. Courtesy HS2
Jackson Pollock in action: diagram showing possible evening peak-hour service pattern on West Coast fast line, flighted to maximise space for slower services (110mph Class 350s shown in green, 125mph Pendolinos in red); each vertical division represents five minutes. The blue lines on the diagram show two Class 350 trains making Fast line calls at Watford, Hemel Hempstead and Berkhamsted. These are the ones for which paths can be found (HS2 proposes three such trains), and even this does not consider whether Slow line paths are available after crossing at Bourne End or Ledburn. The loss of usable capacity north of Ledburn caused by these trains is apparent, as is the irregular pattern of stops at Milton Keynes caused by flighting to maximise the opportunity for slower trains.
18TPH CAPACITY
For HS2, the intention to cater for 17 paths per hour on the Phase One core, plus an 18th as yet unallocated, is maintained. The Business Case makes reference to a University of Birmingham paper that sounds lukewarm about that intensity of usage, but by praying automatic train operation (ATO) in aid claims it is feasible. Some of the assumptions in this paper are also said to be conservative, and to my reading it contains a big logical flaw in its treatment of the effect on headway of diverging trains.
But both arithmetic and modelling show that, with European Train Control System (ETCS), a technical headway of around 120 seconds, increasing locally at features such as neutral sections to around 130 seconds, is feasible even at 360km/h (and remember that the normal operating speed will be only 330km/h).
However, the case with high-speed running is irrelevant. Where things become tricky is in slow speed areas around stations such as Old Oak and Euston, where the binding headways of up to 150 seconds are found. The signalling scheme for an intensively used station must be the product of co-operation and iteration between signal engineer, permanent way engineer and operator, whatever the top speed of trains on plain line in open country might be. What HS2 needs is signalling designed to the good old Southern Region philosophy of using signals to keep trains moving when all is well, not just stopping them when it isn’t!
But presumably not by coincidence, 18 trains per hour (tph) with 150-second technical headways precisely equates to the UIC 75% guideline for usage of capacity (that’s the 75% that Lord Berkeley for reasons known only to himself applied twice over). So as 18tph passes the theoretical tests, move on to consider the practicalities. The main fear is of late handovers from the conventional network, but with all trains having identical behaviour on the HS2 core route the key risk – of a fast train falling behind a stopper – does not arise. A late train, however late, just squeezes in and the delay decays over the next three or four trains. HS2 also has options for ‘graceful degradation’ that the existing railway does not, such as turning short at Old Oak Common to recover time or, by planning the service into 10 platforms at Euston, keeping the eleventh free for a hot standby or to accept the late running train it replaces.
Full service: HS2’s plan for 17tph once the full ‘Y’ is open. Courtesy HS2
HOT STANDBYS
To give extra assurance over workings at Euston, an off-peak service with a couple fewer trains might be devised, perhaps by working Newcastle trains by extending the portions detached at East Midlands from Sheffield trains, currently planned to run only to Leeds or York. But surely to goodness we are not going to spend billions on new infrastructure only to not use it to the full because we don’t trust the railway we already have – the railway we would have to continue to over-use in the absence of HS2! The message must be to make it work, by hot standbys, contingency plans or whatever, and if that also means spending money on the conventional network to minimise late running, then do that, to the good of everyone, not just high-speed travellers.
If anyone still wants to descope Euston and pack the trains into fewer platforms, the analysis makes clear how much of the claimed benefit of HS2 comes from improved reliability, so there is a lot to be lost by penny-pinching and building in a constraint that can never be eased at a location so critical as the London terminus. Yet again, there is no substitute for doing Euston properly, and as soon as possible.
West Coast ‘350’: unit No 350241 at Old Linslade on 1 June 2020. Fraser Pithie
FARES
Another interesting note is that the economics have been worked out on the basis fares will not change compared with now. This, one in the eye for those who claim no-one will be able to afford to use HS2, is logical, as unlike Kent and HS1, for many places HS2 will be the only service rather than being a high-speed option. The only route for which a premium might be justified is Euston – Birmingham, and the market might stand it, but why build a new railway only to price people off it?
In terms of the detailed design of the railway and its train service, this Business Case is ‘the end of the beginning’. Even so, it is a matter for some concern that a classic service should be tabled that even a little ‘back of envelope’ work shows to be so unrealistic operationally. In so far as it has a commercial rationale, it appears to be to enhance commuting services for mature markets at the southern extremity of the WCML, but this comes at a heavy opportunity cost in terms of improved services for major growth areas at Leighton Buzzard and north thereof. The serious debate on train services, especially on the classic network after HS2, must now start in earnest.
William Barter
William Barter is a specialist in railway operations and planning. His career started in operations management positions with British Rail, before moving on to consultancy in the UK and overseas. He is a Fellow of the Institution of Railway Operators, and tutors students on the IRO degree course modules for Operational Planning and Economics.
