Return to Archive -by date - by topic - 2000 Archive.

INFORMED SOURCES April 2000

Vehicle approval gets political

Prescott puffed, Morton huffed but Corbett wouldn't be pressurised on safety cases

Shortly after the Southern Region's Bournemouth Electrification was completed in 196??(help), an engineering chum of mine was in the cab of an electric multiple unit keeping an eye on the new railway. As the train approached a series of signals the driver asked, ‘What colour would you like, guv?'

‘What do you mean', said my chum?

‘Watch this', said the driver. And as he manipulated the power controller, the signals ahead changed their aspects.

Eek. And that was with English Electric electro- mechanical resistance control. Subsequently, as higher tech electrical kit came in, similar interference with signalling and telecommunications was noted over the years.

Defense

Of course, signal engineers have developed ways of getting round such problems, but if you have been reading this column for the past decade you will know that the arrival of three phase drives, which depend on feeding variable frequencies to he traction motors, represented a credible threat to track circuits, particularly the simple 50Hz ac track circuits south of the Thames. Something like a Networker or a Class 92 locomotive returns thousands of amps to earth via the rails and a fault could produce just the frequency to trigger a track circuit, causing wrong aside failure.

You will also know that, unlike continental signalling practice, Britain does not have sequential proving. This technique ensures that if a train leaves a track section, the next section has to be occupied before the first one is cleared. In Britain trains can vanish unnoticed.

More BoS

So, as faithfully reported in these columns, train builders' Brains on Sticks (BoS) have laboured long and hard to demonstrate to their Railtrack peers that under credible fault conditions the new three phase drive trains will not upset track circuits. And as the table shows, it can be done, starting with Holec (now Traxis) on AC and most recently including Alstom with SWT's Class 458 Juniper on DC

Juniper joy

To digress, getting the first Intermediate certificate on DC, clearing the Class 458 to carry passengers (‘I passenger' in the trade) is a huge achievement by Alstom, their consultants AEA Technology Rail, their partners SWT and the train's owners Porterbrook. And it was one of those spirit lifting days when I stood at Waterloo and saw Class 468 units Nos 5 & 6 arrive on their inaugural run – packed with half term traffic.

To digress even further it was even more encouraging to learn, as I rode to Alton and back in Unit 6, that since arriving from Alstom's Washwood Heath plant, its only previous mileage was a quick trip down the Shepperton brach to check that the frequency ‘footprint' was tickety boo. So, new train, into public service with less than 100 miles on the clock.

Final digression, the train was not fitted with steps, because of ongoing gauging issues with Railtrack. So congratulations to Her Majesty's Railway Inspectorate who signed off the train to enter service on the rational basis that a Class 458 temporarily without steps was a darned sight safer that a Mk 1 slam door EMU.

End of digression. By the time you read this, the Class 460 Gatwick Express unit, also DC, should have got its I Passenger Certificate and, if Alstom can sort out an overhead line resonance issue, the ink could still be drying on the I Passenger for Scotrail's Class 334 AC Juniper. Resonance, which can affect the electrification system or other trains, is a reminder that there is more to electrical safety cases than signalling interference

Derby depression

But over on the other side of the M1 things are not so rosy, as Deputy Prime Minister John Prescott discovered on another of his tours of the train builders on 11 February.

At Alstom they took their harangue about delivery and quality on the chin and showed the BUFF their rainbow production lines. But at Adtranz they showed him 46 Class 457 trains for LTS Rail still in store and whinged about the iniquity of Railtrack's approval process.

And Prescott went back and spoke to the Shadow Strategic Rail Authority. And on Monday 14 February there was a meeting between Railtrack and Adtranz on approval, with David Humphrey, the SSRA's traction and rolling stock wonk, in attendance.

And after the meeting he reported back to SSRA Chairman Sir Alastair Morton about the lack of progress and the next day Sir Alastair wrote a letter to Railtrack Chairman Sir Philip Beck to whom Railtrack's Safety & Standards Directorate reports.

And this is what Sir Alastair said.

Morton's letter

Now this is the sort of letter you write in the heat of the moment, feel better for it, put it in a draw overnight and throw it away next morning. But Sir Alastair posted it and a copy arrived on my desk.

Appalled

And I was appalled. Yes, I have been critical of acceptance, but Railtrack has tried hard to improve its process in response to criticism from train builders and operators. And since the process came in trains have got safety cases or had safety cases extended.

Clearly, the SSRA doesn't understand this, because when I spoke to Sir Alastair about the letter, I could not get him to accept that in February 2000 we did not have an acceptance problem, we had an Adtranz problem. I mean, if you can extend the safety case for the Class 92 loco – designed before the new process was invented - up the West Coast Main Line Museum of antique track circuits and back down the East Coast Main Line via Carstairs to Doncaster, and without changing the hardware, why can't Adtranz get a route specific AC safety case for LTS Rail, a short route with only three types of track circuit?

And if Alstom can get a DC safety case after 15 months of testing, with few restrictions on test running in that time, why is the Class 375 still restricted to testing under total possession on Connex South Eastern?

Well, one answer is that Alstom came up with a cunning plan. Three phase drive Networkers can run because they are fitted with an Interference Current Monitoring Unit (ICMU). This monitors the traction return current for the 50Hz frequency that could interfere with track circuits and if the current at this frequency exceeds a specified level, shuts the train down.

So the Class 458, which has internal frequency monitoring and doesn't need a separate ICMU, began tests with an ICMU fitted. This was a bit of a bugger, because ICMUs are hair trigger, but it meant that the class 458 was as safe as a Networker and was soon able to be tested without the need for a possession.

Misunderstandings

But back to the letter, which I find full of misunderstandings. Take the reference to the process being ‘rethought'.

Railtrack has been rethinking and refining the process for around five years now. A fortnight after the letter was sent, Issue 3 of the Yellow Book was published. This is the document which sets down a process for Engineering Safety Management.

If you change something, like introduce a new train, you have to demonstrate to Railtrack that you are not introducing additional risk to the network. The Yellow Book describes one proven process for achieving this.

Issue 2 was, if anything too comprehensive, its four volumes including a two inch think ring binder. Users regarded it as over detailed and too prescriptive.

So Railtrack set up a 15 strong steering group with representatives from train operators, train and signalling manufacturers and the HMRI. A parallel User Group generated feedback on experience with Issue 2.

As a result, issue 3 fits ‘Fundamentals' and ‘Guidance' into a half inch thick volume which was published on 29 February. Over 900 people have been through the two day training course on Issue 2 and there are half day conversion courses for Issue 3.

So the process is clear and those on the receiving end have helped improve it. But what about the ‘ultimately vain attempt to eliminate risk from the system' instead of using ALARP? Not so, in my view. What you have to do is identify risk in a new train, and then demonstrate how you control or ameliorate it to acceptable levels.

This is a big advance on the original BR Specification where you had to demonstrate that your traction package would not produce nasty frequencies more than once in half a million years. Now that really was bonkers (this column passim)

Independent assessment

Nor is acceptance just a case of Railtrack saying ‘yes' or ‘no'. Each manufacture has an Independent Safety Assessor (ISA).

His job is to double check the safety case proposal which goes to Railtrack's Systems Review Panel (SRP) which is responsible for acceptance. And on occasions, the ISA has been unwilling to support a submission before the SRP.

And, spare a thought for the BoS on the SRP who sign off a train as safe to carry passengers. Back in the days when we had engineers of stature, they would sign off new traction on their own responsibility. But in those days you did not risk prosecution or media persecution.

Imagine you get it wrong today and a fatal accident ensues. The next thing is your face on the front of the Daily Wail as ‘The man who signed off killer trains'.

Support needed

When I spoke to Sir Alastair, he emphasised that his real concern was not with the levels of risk behind the SRP's decision but with Railtrack's perceived lack of support for manufacturers during the acceptance process. The SSRA wants to see Railtrack being ‘more proactive' in helping manufacturers produce acceptable safety cases.

But, hang on, as a result of earlier criticism from the industry, each manufacturer already has a mentor tasked with doing just this. The mentor also sits on the System Review Panel, and can thus provide guidance on where a submission is falling short and what the SRP needs to see.

However, I suspect that what the SSRA would like is for Railtrack to tell a manufacturer what to do to it's kit to get acceptance. Hang on, the guys on the SRP may be BoS, but they are concerned with what the outputs from the traction package do to the infrastructure. I doubt whether they have the traction engineering expertise to tell Adtranz or whoever, ‘If you put a 3 megohm resistance in the IGBT firing circuit that should do the trick', or ‘Line 121023 of your software has an ELSE when it should be an OR'.

Railtrack could also be more supportive in organising and managing possessions for testing more effectively. According to an SSRA Informed Source ‘We want Railtrack and Adtranz to work together 24 hours a day to get these fleets of trains at present standing idle into service as fast as can be achieved safely'.

Gerald replies

It was Railtrack Chief Executive Gerald Corbett who responded to Sir Alastair's blast. Why Corbett and not Beck? Well the SRP is part of Railtrack Line not S&SD, which reports to Beck.

Now, remember that this letter arrived as Railtrack was under the cosh from Prescott over safety, see below. And Corbett had been taking continuing flak over safety since Ladbroke Grove.

Not surprisingly, in these circumstances, his reply was robust. The SSRA thought it a bit ‘uncompromising' Judge for yourself.

Gerald Corbett's reply to Mike Grant

That Gerald Corbett is well informed, or well advised, comes in that final ‘or lack of it'. This referred to problems with tests that very week.

Private enterprise

So there we are then. For sure, the acceptance process is not perfect, but just because one supplier has problems does not mean it is fatally flawed and it is being improved.

With its old railway background, the SSRA seems to be living in the days when British Rail Engineering built trains to BR's spec and it something when wrong everyone pitched in to help for the good of the railway.

It ain't like that any more.

In the private sector a private operator has a choice of private builder but caveat emptor now applies. And if a supplier fails to deliver, well, it was your call, you got it wrong and you'll know better next time.

Equally, get it right and you should reap the rewards. Thus SWT has spent around £750,000 on the Class 458 acceptance testing while Alstom must be looking at £15million for its Juniper safety case programme.

These are the entry fees to the new train market but the SSRA don't seem to have got the message that the old railway is dead. When the Class 458 got its I Passenger, the reported reaction from the SSRA was not ‘Thank goodness, at least someone will be able to supply Mk 1 stock replacement units in time'. No, they asked Alstom to share its expensively won expertise to help out rival manufacturers having trouble with the acceptance process.

There are times when I despair.

++++

It can be done

Electrical safety cases accepted to date

Notes

* Under current process

a) West Coast main line throughout plus East Coast Main Line Carstairs to Doncaster in addition to existing DC Channel Tunnel routes

b) Paddington-Airport Junction

c) In addition to DC

d) East Coast Main Line in addition to DC Channel tunnel routes.

Continues.........