IEP EMC hits GWR, LNER and NR

Acronym heaven - and there’s another pink flamingo in there somewhere

Informed Sources

IET on test: unit Nos 800002 and 800001 on working 5X80, the 08.35 Old Dalby to North Pole depot via Reading, at Twyford on 29 July 2016.
Ken Brunt

On 9 December 2003, I made my way to Ilford depot to see the first Hitachi ‘V-train’, a converted Class 310 incorporating a threephase drive traction package. My hosts were Hitachi’s then Managing Director Europe and the company’s Business Development Director, one Alastair Dormer who, 15 years later, merits the famous remark made about Sir Christopher Wren: ‘si monumentum requires, circumspice’.

At that time, getting a safety case for alternating current (AC) traction was not easy. Having been looking at the UK market since 1999, Hitachi had realised that the infrastructure was ‘very demanding’. And while the company was confident that the test would show its traction technology to be 100% compatible, it was clear that this had to be demonstrated, with ‘Electro Magnetic Compatibility (EMC) a key issue’.

With V-train Hitachi would both demonstrate compatibility and acquire the measured data needed for the safety case process. Not just the traction package would be monitored and recorded. Noise, vibration and UK track condition data would be observed.

A-TRAIN

Installed in the Class 310 was a version of the Hitachi A-Train traction package, uprated for the proposed Channel Tunnel Rail Link domestic services, then out to tender. Hitachi would be announced as preferred bidder for what is now the Class 395 in October 2004. In addition to the actual drive, the Class 310 also incorporated the transformer, rheostatic brake resistor and auxiliary inverter.

Testing under the 25kV AC overhead line was due to start that month, followed by 750V DC in February 2004. The kit in the V-train had already been subject to extensive development and testing, including running a prototype traction package to run on a dynamic test bed, replicating the outputs from UK traction substations. By the time of my visit, around 25% of the safety case for the new UK A-Train was already complete.

QUALIFIED

This trip down memory lane emphasises that Hitachi knows more than a bit about the UK rail infrastructure. Despite this, the 800 Series Intercity Express Trains (IET), the train formerly known as IEP, have been bedevilled by EMC issues.

Test running began when the overhead line equipment on the 16-mile ‘test track’ between Reading and Didcot was energised in July 2016. And promptly stopped due to electrical interference problems.

I assumed it was the infrastructure, but in a visit to Hitachi’s Newton Aycliffe plant in December that year (‘Informed Sources’, February 2017), the company confirmed that type testing was still progressing, with Hitachi working in parallel with Network Rail through electrical interference issues. ‘We just have to come up with a solution, don’t we’, Hitachi Rail Europe Chief Executive Karen Boswell told me.

With DfT desperate to get its flagship project into service by the October 2017 target, the GWR Class 800 fleet was given a Qualified Type Acceptance Certificate. Electric operation was restricted to Paddington-Didcot.

But if interference had raised concerns on the GWML, it became a potential show stopper when testing in electric mode began on the East Coast main line (ECML). It was discovered that when an IET entered the York Integrated Electronic Control Centre (IECC) area, the Solid State Interlocking (SSI) Data Link Modules, which drive the points and signals, began flashing up faults, including an occasional failure. As a result, IETs have been banned from running under electric traction north of Colton Junction.

QUALIFIED ACCEPTANCE

Any Qualified Type Acceptance Certificate issued in respect of a Set shall list in detail the outstanding conditions to be satisfied and tasks to be performed by the TSP in order for a Type Acceptance Certificate for that Set to be issued (the Type Acceptance Preconditions), in accordance with a timetable set out in the Qualified Type Acceptance Certificate, which must in any event be satisfied before any Final Acceptance Certificate may be issued in respect of any Set of that Type. Great Western IEP Master Availability & Reliability Agreement

PENDOLINO PROBLEM

Alstom’s Class 390 has an IGBT traction system. At the time it was introduced it was assumed that because the switching frequency of 1.8kHz (1Hz – 1 cycle per second; mains electricity is 50Hz) was much higher than the 400Hz of GTO drives, interference would not be an issue.

Alstom had used various techniques to minimise harmonic currents, although there were a few Amps still drawn at 11kHz and 22kHz. These were considered well above railway telecommunications and track circuit frequencies.

False expectations: some signalling and telecommunications systems were found to operate in the 10-50kHz range. It didn’t help that the relevant standard for EMC (EN 50121) was still in draft and didn’t reflect the systems on the WCML.

During early Pendolino testing in 2001 technicians at Birmingham New Street power signal box detected intermittent problems with the SSI controlling Proof House Junction. Monitoring showed that there were serious problems with the data link.

These started as a Class 390 approached and departed New Street. They continued until the train reached Stechford and switched to another OHLE feeder section. Tests showed that the voltage induced in the Data Link cable was not that different to that generated by other trains. However, the frequency range was different and the amplitude much higher. Further work showed that a DLM was generating and transmitting anomalous pulses as a result of the common mode interference, which corrupted the SSI telegrams.

ANOMALOUS PULSES

Some head scratching concluded that while the DLM had an internal transformer, which should block the onward transmission of the interference, defects in its design allowed the anomalous pulses to be generated. Modelling showed that adding an adjacent DLIT would solve the problem. This was fitted to allow Class 390 testing to continue while the root cause was found.

What was going on inside the DLM stretches even my mechanical engineer’s brain so I am not going into it, save to say it was connected to the design of the internal transformer. Today, the designer of the unit agrees, but now thinks that the investigation fingered the wrong mechanism.

Four types of DLM from two companies were tested. While their susceptibility to interference varied, the worst case frequency for all of them was around 20kHz.

While a DLIT was shown to block common mode interference, it was not a quick or simple universal solution. DLITs would have to be shown on signalling diagrams and signalling works testing would apply to installations.

FLAMINGO AHOY!

At this point a pink flamingo flaps lazily overhead. Jarvis (remember them?), who had taken over infrastructure maintenance on the ECML, had become concerned that the lightning protection in the SSI system might be damaged by induced voltages in the data link cables. One of the company’s engineers developed a Surge Protection Unit that also incorporated an isolation transformer.

This was a direct replacement for the existing SSI surge protector and when installed at Proof House Junction was so successful that after the tests it was left in place. However it did have some limitations for wide-scale use. As a result Park Signalling – the railway equivalent of the legendary Lockheed Aircraft ‘Skunk Works’ – was commissioned to produce the final abbreviation in this saga – the Isolating Surge Protection Unit (ISPU).

Development of the ISPU was completed in the summer of 2004 and it went into production. Consultant Lloyds Register produced ‘Application rules for Isolating Surge Protection Units’. As a result the Pendolino timetable was introduced in September 2004 and they all lived happily ever after. Well, not quite, because a cloud of amnesia then settled over the Kingdom of Network Rail.

MORE FLAMINGOS

In June 2018 Class 800 bi-modes started running in electric mode in the York IECC area. And, you’ve guessed it, TFMs started flagging up error messages. The interference was so serious that the initial ban on running north of Darlington became north of Northallerton and then north of Colton Junction.

East Coast test run: No 800101 running northbound on the East Coast route through Morpeth with working 5X05, the 08.36 Doncaster Carr to Edinburgh test run on 9 November 2017. The Voyager in the photo was the 09.00 Glasgow Central to Penzance CrossCountry service.
Bill Welsh

But because of the amnesia cloud, no wise old signal engineer said ‘Hang on a minute, new three-phase traction interfering with SSI, that sounds like a replay of Pendolino on the WCML’. So everyone retreated into their silos, declaring it was someone else’s fault.

Fortunately LNER had now taken over Inter-city East Coast, with Chairman Robin Gisby, a former ECML Route Director, acting as DfT’s representative in the real world. A meeting of all parties was called and, amid much hurling of aspersions, they were required to come up with a solution.

Naturally, the instinctive response was to call in consultants, in this case Ricardo. At a follow-on meeting on 4 September, Ricardo’s report was presented and according to informed sources it drew heavily on the 2004 report into Proof House Junction produced by Network Rail, Atkins, Park Signalling and Bechtel. At the meeting it was agreed, surprise, surprise, that ISPUs were the solution to SSI interference.

Now while the Class 800 is particularly ‘noisy’ in EMC terms, the fact that the DLM has a known weakness means that Network Rail doesn’t have a leg to stand on in this specific case and will have to pay to bring its infrastructure up to scratch.

With legacy SSI in York IECC, Tyneside IECC and the Morpeth, Alnmouth, Tweedmouth and Edinburgh Signalling Centres, there are likely to be multiple locations affected. Presumably work will now start to determine which locations match the criteria for ISPU fitment.

South of Colton Junction, the only SSI are at Peterborough, Hitchin and Finsbury Park, and these use relays for lineside interfaces, so do not suffer from the DML problem. Just in case, this has been confirmed by testing.

Just to round off the SSI story, amnesia was selective. In Scotland, the resignalling for the Edinburgh to Glasgow Improvement Programme had the full house with LDT filters, ISPUs and DLITs all ready for when the electrification went live.

CONTINUING CONCERNS

So end of IET interference issues? Unfortunately not.

When electric traction testing with the Class 800s began on GWML it soon became clear that, despite what I have written in the opening paragraphs, the electric traction package was seriously noisy. The two charts opposite show the results of back-to-back testing against a Class 387.

I expect I will be told that these are different types of train and that I am comparing apples and oranges. In my book, apples and oranges is the argument of last resort when facts are inescapable.

In this case both trains have pantographs, transformers, converters and inverters feeding traction motors. A 10-car Class 800 has a traction rating of 5.4MW compared with 5MW for a 12-car Class 387.

So at best we are comparing the Vitamin C content of citrus fruits. Under the route modernisation programme, GWML has been resignalled with the modern equivalents of SSI – Alstom Smartlock and Siemens Westlock. Train detection is by Thales axle counters and level crossings have closed circuit television (CCTV), both of which involve trackside data links.

AXLE COUNTERS

On the GWML there have been instances of axle counters failing-safe and indicating occupied after a Class 800 has passed. In May, for example, an axle counter on the down main failed, together with that for the parallel section on the adjacent up main.

Axle counters present a different problem when it comes to interference. While they have trackside wires transferring signals from the detector head to the processing unit, they also need power at the head. This means running power and signals down the same wires, which rules out a simple isolation transformer.

Hitachi told the ECML meeting on 4 September that there had been no axle counter failures on the GWML since 11 May. However, Great Western Route sources remain concerned, noting the delay to testing with 800 Series units in electric mode west of Didcot.

COMPLIANCE

Add in interference affecting level crossing CCTV and we are far from out of the woods. While NR has been embarrassed over SSI on the ECML, on a recently-modernised route like the GWML it should be up to the rolling stock manufacturer to prove compatibility with infrastructure under The Railways and Other Guided Transport Systems (Safety) Regulations 2006 (ROGS), which came into force in 2006. ROGS implements the European Railway Safety Directive (2004/49/EC), which aims to establish a common approach to rail safety and support the development of a single market for rail transport services in Europe.

And it remains a fact that the Class 387 and Class 345 have gone into service on Western Route without interference issues while the GWR Class 800 fleet is still running on Qualified Acceptance certificates. Meanwhile, over on the ECML, the King’s Cross throat remodelling will be based on axle counters.

WHAT NEXT?

Well, a lot depends on the politics of Type Acceptance. According to informed sources, DfT is unhappy that the GW trains are continuing to run under Qualified Acceptance certificates with issues still outstanding. It is determined that it won’t make the mistake twice.

Meanwhile, Hitachi is reported as being confident that it can obtain Type Acceptance for LNER to introduce a Leeds service in December, followed by York in February and Newcastle in March next year. How this relates to Network Rail’s programme for immunising the SSI data links is uncertain. I am told there is already a backlog of ISPU orders.

Clearly it is in Hitachi’s interests to get trains fully accepted, because then, even if the infrastructure prevents them from running under electric traction, LNER still has to pay for the diagrams. Note that under diesel power a Class 800 will not keep to time north of Colton Junction. Equally, since procuring and installing ISPUs is going to take time, you can see why it is in DfT’s interests to hold out for full acceptance certificates or nothing.

TIMETABLE

Then again, LNER has to put in its May 2019 timetable bid, which could be a bit tricky if you don’t know whether and how many IETs you will have available and where they can run. The obvious solution is the classic infiltration, with Azumas running in IC125 paths as they are accepted.

However, this lacks the revenue boost of a full-blooded launch and in the words of one informed source would be ‘damned expensive’. Meanwhile time’s winged chariot is on the heels of various IC125 operators including LNER. The day this issue is published there are just 460 days until all passenger rail vehicles must be accessible. Any delay to IET deliveries could see LNER requiring a derogation to maintain services.

Meanwhile, Automatic Power Change Over (APCO) balises (I thought we’d had the last abbreviation – Ed) were installed near Temple Hirst Junction in September and are scheduled to remain in place until testing with Azumas ends on 2 December. An APCO balise tells the IET when to switch, automatically, from diesel to electric traction or vice versa and I understand this is one of the outstanding issues in the GW Qualified Acceptance list. As is its close relative in the balise world, ASDO (Automatic Selective Door Opening). (That is definitely the last abbreviation – Ed.)

HOSTAGE TO FORTUNE

‘The Intercity Express Programme (IEP) Invitation to Tender was published by the Department in 2007. The Invitation to Tender included a technical specification, the purpose of which was to ensure compatibility between the train and infrastructure, by providing railway infrastructure information to support the development of the train design. Therefore Hitachi are required to work with Network Rail to ensure that the train is compatible with the existing East Coast main line (ECML) infrastructure. The signalling system on the ECML has not been upgraded ahead of the introduction of the Intercity Express Programme trains.’ Jo Johnson, Rail Minister, Written Parliamentary Answer, 13 September 2018