Thameslink and the NAO: Part 2 – Wherefore Art Thou Siemens Rolling Stock?

In part one we looked at the verdict of the National Audit Office report on Thameslink and more specifically on the infrastructure work done so far. In part two we look in detail at the comments relating to train procurement.

Here, one does have to get far into the report to realise that there is a big problem that needs to be sorted out by the DfT. A telling paragraph describing the problem is as follows:

11 The award of the contract to buy new trains is currently delayed by more than three years. The Department decided in March 2008 to let a complex PFI contract to one supplier to design, build, finance and maintain the new trains. We cannot comment in detail on the reasons for the delay to the procurement until the contract is awarded. However, the delays raise questions about whether the Department underestimated the scale of work, time and skills and capacity it required to negotiate a complex PFI deal.

As far as impacting on the Thameslink Programme, the next paragraph states:

The delay in letting the contract for the new trains:

  • Adds logistical complexity to the infrastructure project. In particular, there are risks around accommodating the train design with some elements of the infrastructure project.
  • Adds complexity to the process of letting a new franchise. Uncertainty around when the train contract will be let makes it more difficult to determine appropriate terms for the franchise agreement.
  • Raises questions about the feasibility of delivering the whole programme by 2018.

The Department has not yet fully mapped out the programme’s critical path based on a revised timetable for delivering the new trains to determine whether 2018 is still feasible or how much contingency is left in the timetable. The Department has been working with its industry partners to do this but cannot complete this until it has let the contract for the new trains.

It does not end there…

13 The delays also affect the Department’s plans for the electrification of other parts of the rail network. The delays mean that the Department is currently buying additional trains to meet short-term demand on newly electrified routes elsewhere in the country. The additional procurements raise questions around who bears the risk that the trains will still be needed when the Thameslink procurement is completed, and the Department’s role in securing and managing the allocation of trains within the network. We intend to return to this subject in future work…

This clearly is a major issue because part 3 of the report is entitled “The impact of the delays to the new trains” and it doesn’t make pleasant reading.

The Revised Timetable

Within part 3 is a crucial table that contains three columns of dates. The first set of dates (column 2) is the timetable as announced at the time of issuing the Invitations To Tender (ITT) for the trains. The second set of dates (centre column) shows the revised schedule set in 2010 announced ostensibly as a result of the spending review but in part, at least, because it was realised that the original schedule was not realistic. The third set of dates (column 4) list the dates when the critical milestones mentioned were actually met.

Key Procurement Dates

Critical Procurement Dates and how they haven’t been met. Click to enlarge.

The DfT keep saying that it is still expected that all the trains will be delivered in time for the planned opening in December 2018. Given that is over five years away that sounds reassuring. The problem is that plans outside Thameslink originally relied on the Thameslink trains being delivered much earlier. According to the original schedule (second column) the first of the batch should have been in service by now and enabled some Southern Class 377 on loan to First Capital Connect (the current franchise holders for Thameslink) to be returned to strengthen Southern services.

It was, of course, realised that the delays to the Thameslink stock were going to cause problems elsewhere. This has been resolved by ordering other replacement trains and we will look at that in detail in a moment. Meanwhile, even if the revised delivery target (middle column) had been met, the first trains would not have arrived until July 2015. This would have created a problem because Thameslink trains will be unable to call at London Bridge from December 2014. The alternative route via Herne Hill is approximately nine minutes slower so just to maintain the Thameslink service during this period would require extra trains. On top of that, the infrastructure is quite capable of taking 12 car trains on the Brighton route. It would be a sad reflection on a scheme that has the objective of increasing capacity if the process of implementing it caused capacity to actually be reduced. Yet due to a lack of rolling stock this is what would happen, unless some more trains appeared from somewhere. Fortunately this appears to be the plan.

The Risk of a delayed project due to no new trains

What is more worrying still is that if the trains are not ordered in time it might be the case that the project cannot complete on time. The revised timetable (middle column) allowed a generous 3 years 9 months to deliver the first train and a further three and a half years to deliver the final train. The problem is that although Siemens were announced as preferred bidder in June 2011 the contract still has not been signed.

Even if it were to be signed next month (July 2013) this delay means that the first train would have to be delivered in two years time (July 2015) to get back on the revised schedule. As we are not talking about an existing product that comes off an existing production line, but a train that doesn’t yet exist, this would appear to be very aggressive delivery schedule indeed.

Don’t Know, Can’t Ask

Rather frustratingly, the report suggests from one of the paragraphs quoted earlier that the DfT will not be able to know if the project can be completed on time until  the contract is signed. The suggestion is that the questions they need to ask other contractors and Network Rail cannot be asked until there is a contract in place, due to commercial confidentiality.

Playing Catch-up

If the delivery of the first train does not take place until after July 2015 then the only way the project can get back on schedule is if trains are delivered at a faster rate than planned. Three and a half years might sound like a long time to deliver a fleet of trains but we are talking about a huge order expected to consist of around 1200 hi-tech carriages built to a design that hasn’t been tested in the real world.

One only has to look at S stock deliveries, or even the Victoria Line 2009 stock deliveries, to realise that one carriage a day often represents a typical rate of train production. As a further complication, delivery and acceptance (usually after an agreed accumulation of miles) are different things and if any problems occur during the delivery phase it may be necessary to recall vehicles already delivered and rectify them before restarting new deliveries. Sending trains built by Bombardier back to Derby is bad enough, but at least there is a relatively available route to do this on an ad hoc basis. Sending them back to where Siemens will assemble the trains, which according to the BBC is South Tyneside but in reality it is almost certainly Germany would probably not be practical and may be another reason to have the depots ready long before the first train arrives.

Even if you could produce the trains at a higher rate than planned this does not solve everything. It is important to get a few trains early on just to test compatibility and the report makes much of this. Of particular concern is the ATO signalling. Of course nowadays signalling has quite a large train borne element and for ATO this is especially true. Those responsible for project delivery will be looking to have as long as possible to thoroughly test the interfaces. The only good news here is that Siemens recently bought out the signalling supplier, Invensys Rail, so hopefully there won’t be inter-contractual disputes. One wonders how much the Siemens Thameslink order influenced the decision by Siemens to buy Invensys.

Of course things can be done in mitigation to try to cut out as much delay as possible. It is known that Siemens have gone ahead with the development of a new bogie even though they haven’t got the final confirmed order yet. The DfT and Siemens have also entered into an Advance Works Agreement which enables Siemens to commence construction of advance works at the two depots – Hornsey and Three Bridges. Obviously it would be unfortunate if Siemens managed to produce trains in record time only to find they couldn’t be delivered because the depots were not complete.

We are sorry for the delay. The reason is confidential.

As stated earlier the exact reason for the delay in signing is not clear and will remain confidential until agreement is reached – assuming it is reached one day. Initially it was believed that the problem was that, due to the worldwide recession, Siemens had a problem borrowing money at the rate they had budgeted for in order to finance production for this large contract. This might have been plausible at the time, but with money now available at low rates of interest this seems less likely. On the other hand, the government has made an exception for Crossrail and allowed it to buy the trains it will order. In other words they will be publicly financed. This is to ensure that Crossrail does not suffer from the delays that Thameslink does. This would seem to only make sense if the Thameslink problems caused by issues involving the private financing of the order. Notwithstanding the revised financial arrangements, the Thameslink rolling stock delay must be very frustrating for Crossrail, who were probably originally hoping that could see the new trains in action before finally committing to the Crossrail order.

The NAO report rather suggests that, in trying to do an admirable thing and issue a tightly-specified Design, Build, Finance and Maintain contract, the DfT has seriously underestimated the complexity of such a task.

The belief is that neither side can afford to back out of this contract. One would like to believe this, but a worrying question would be what would happen if an agreement just could not be reached in time – or at all? We really would be in uncharted territory.

What if not all of the trains are delivered on time?

We are led to believe that it will  be possible to run non-ATO-equipped trains in the ATO section. There will be conventional signals and, presumably, TPWS will be active. If true, this must be a saving grace. Presumably if Thameslink were able to hang on to existing trains then they could operate some kind of enhanced service from December 2018 even if not all the new trains were delivered.

If the December 2018 implementation relied on some non-ATO trains then it would seem likely that there will be only 20tph – 22tph at most. The problem with going for 24tph without full ATO is that if it doesn’t work there has to be an embarrassing climbdown and a new less-challenging timetable introduced. If, on the other hand, it does work, then serious questions are going to be asked about the necessity of specifying ATO the first place. From Network Rail’s point of view, assuming we ever reached this stage, the best outcome would be that it is realised that 24tph works, just, but it is clear that with ATO there is the advantage of being able to recover much more quickly from a disrupted service and so specifying ATO can be justified.

The Replacement trains

The NAO report very helpfully records in a table trains all the units that either have been purchased or are due to be purchased as a result of Thameslink rolling stock delays. A brief description of the reason they were needed is given and the cost to the department caused by the need to order these trains as a result of a delay to the original Thameslink order. So the last column is the cost of the delay – not the cost of the trains. The implication is that had the original deal gone though when intended the DfT or TOC would have got a more competitive deal with lower operating and leasing costs.

Additional Train Procurements

Carriages ordered as a consequence of Thameslink Delays. The final column headed “cost” is the cost attributable to the procurement delays – not the cost of the carriages themselves. Click to enlarge.

How much does a railway carriage cost?

From the information provided in the table and the original press releases issued by Southern – here and here – we can get a very good idea of an approximate figure of what a railway carriage costs. The figures must be treated with some caution as it is impossible to tell if we are comparing like with like. For example, the presumption made is that all figures quoted are capital costs and don’t included a maintenance element.

Taking the Siemens trains first, the figure of £1.6 billion has been stated as the approximate value of the contract. This covers the construction of two depots. Let us assume £50 million to build a depot – which is probably on the low side. It is presumed that approximately 1200 carriages will need to be built to have the necessary number available at any time as specified in the contract. So 1200 carriages costing a total of £1.5 billion means each carriage costs around £1.25 million. This compares almost exactly with the cost of an ‘S’ stock carriage which is shorter, but is highly customised and also comes from a large order.

Taking the most recent Southern order, because this has closer to current day prices and because the figures are simpler, we find 40 Bombardier class 377 carriages will cost £60 million. So each class 377 carriage, in a small production run admittedly, cost £1.5 million. IF these had been the not-yet-existing Siemens carriages, the cost for 40 of them would have been £50 million. The difference in cost, £10 million, probably not coincidentally, is the exact figure quoted in the right hand column as the extra cost as a result of the delay.

Ten million pounds may not sound too bad as an extra cost due to delay in the overall scheme of things. But this is just for the second (smaller) order of substitute trains. Taken with the larger order the delay has already cost £50 million. There will be a further order for which the additional cost is “not yet determined” but is likely to be around £40 million. So that is an extra £90 million due to the delays in ordering rolling stock, and this is not even what the NAO is most concerned about.

A canny collusion?

The orders referenced above originated from Southern. What is strange is that it is Southern that is placing the order – not the DfT. There appears to be some kind of collusion here and for the first two orders this is probably to avoid the open tendering rules in force at the time. These would have applied if the DfT order the carriages instead of Southern. It is believed that the original order was made in December before new rules for tendering came into force in the following year, and the second order was exempt from these rules because it was merely exercising an option present in the contract for the first order. Quite why Southern and not the DfT is placing the third order though is not at all clear.

On the third time round things will have to be done properly according to new rules. So there will be a specification and train builder invited to quote. In this modern world we live in it is not possible for Southern or the DfT to decide that, with very good reason, what they want is a few more class 377 trains and invite Bombardier to put a quote in for that. It has to go out to competitive tender and for that there has to be a specification that specifies what is required.

The other interesting item in the specification for the proposed third round of vehicles is a requirement to be able to run at 110 mph when using the overhead 25kV power supply. This is an advantage to Siemens who can already do this with their class 350/1 trains. It looks like an overlooked benefit of the Thameslink programme may be 110 mph running on the lines lines north of St Pancras – both to Bedford and on the East Coast Main Line. Of course, if ever the justification were made for re-electrification between East Croydon and Preston Park (near Brighton) at 25KV as has been investigated in some detail then with these trains a significant improvement in journey times can be made if the linespeed were to be increased from 90 to 110 mph. It does seem as 110 mph maximum speed will be the new norm for electrical multiple units where achieving such a speed would be worthwhile.

What will the new trains be used for?

The first order was to replace trains which were loaned to First Capital Connect and, due to the delay, Southern are unlikely to see them again for a few more years. At the time of the announcement the reason for the second order was very vague, but the NAO states that this is simply to beef up existing Southern services. This makes sense with longer trains planned for some routes from December 2013. The third order, due to be made soon, is stated as needed to support Thameslink in the short term. Whether this means we shall see 10 car units on Thameslink for a period, or whether these will be substituted for class 377s on Southern in order to release them for lengthening Thameslink trains, or indeed whether they will be used to provide additional London Bridge – Brighton services while Thameslink cannot call at London Bridge, is not at all clear.

An obvious concern is: what happens to the surplus units in the third tranche once the complete Siemens order is delivered? After that they will be used to “provide electric trains for newly electrified routes”. We presume that means routes electrified as a result of the Great Western electrification scheme and possibly others in the North of England. These are the schemes that would have originally have had the cascaded Thameslink stock – presumably after thorough refurbishment.

One further concern, no so obvious to everyone, is that these trains will be ordered for an ultimate purpose but there will initially be no commitment for any train operating company to use them for this purpose. Suppose, for example, that the trains were, for some reason, found to be unsuitable for some future use or that not all of them were needed. Basically there is a risk that the trains will end up in a siding and unused and someone needs to carry that risk. Needless to say the NAO did not miss this point. The report states:

The additional procurements raise questions around who bears the risk that the trains will still be needed when the Thameslink procurement is completed, and the Department’s role in securing and managing the allocation of trains within the network.

Damned with Faint Praise

If some credit could be found for the DfT in this sorry state of affairs, then it is that they appear to have a back-up plan – and arguably quite a good one at that. One almost hesitates to raise the point, but the worry is – what if delivery of the trains in the back-up plan  is delayed for any reason? Or even what if the cannot finalise the contract for third tranche of carriages in time? This must be a possibility given that it involves a new specification, whereas the first and second tranche were merely a repeat order for some existing stock. After all, it is the inability to finalise a contract that has gotten the DfT into this mess in the first place.

In our third and final part we will look at the consequences to the problems of franchising in the context of Thameslink and how the National Audit Office thinks that the Thameslink Programme should be judged.

Written by Pedantic of Purley