Return to Archive -by date - by topic.

INFORMED SOURCES April 2004

Still chasing the unforgiving minute

Reliability – sadly far from boring

Mr Editor Abbott, who has his finger on the readership's pulse, reckoned that the comparisons of fleet reliability in the February Informed Sources would generate a lot of interest – and so it proved. Not least with a page of letters in Forum, including a sharp rap over the knuckles from ATOC for ‘using the information in a misleading way'.

Hmm, well, as you know, this column is the home of the cold numbers, because the cold numbers are all we have to hand much of the time. And the National Fleet Reliability Improvement Programme (NFRIP) figures, however flawed, do tell those of us paying for the railway what we are getting for our money in broad terms.

Of course they have to be used sensibly, rather than to nit-pick. But ss I said in the original article, if a fleet's performance is rubbish compared to its peer group average, all you have to do is double the miles per casualty (MPC) and if it is still rubbish, then there is clearly something wrong.

So, NFRIP figures will continue to be discussed in this column because, while a blunt instrument, they bludgeon through all the contractual sophistry. What is contractual sophistry, you ask?

Well, take the Class 458 (please). Builder Alstom and owner Porterbrook have been saying for over a year that the ‘contractual' reliability is around double the NFRIP figure.

Take the Class 333, According to my old chum David Wilson of Siemens in last month's Forum, the Class 333 reliability figures agreed with Arriva for September, October and November 2003 are as shown in Table 1. As with NFRIP, a technical casualty is defined as a cancellation or delay of 5 minutes or more.

Table 1 Class 333 performance

Source: Siemens

Reliability decline

Well, at risk of making the entire Siemens board choke on their lunchtime <ital>bratwurst<ital>, here are the NFRIP figures for three phase drive electric multiple units for Period 10 of the current financial year (7 December 2003-3 January 2004). The NFRIP Group started collating data in Period 10 2002/03 so we now have a true Moving Annual Average (MAA)

However, unlike the table in the February column I have ranked the trains by their Period 10 figures rather than the rolling average. From which we see that ‘my' local Class 365s and the c2c Electrostars are now even more superior.

That's the good news. But look down the table and see how many other ‘spot' figures are better than the moving annual average. The top three are significantly better, the Gatwick Express Class 460s are about 20% worse and the fifth ranked South Central Class 375s/7s are about 30% better.

All the rest were worse than their MAA – which is interesting and possibly depressing. In fact, almost certainly depressing because the shortfalls on the MAA are mostly into double figures. And the South Eastern Class 365 has the biggest proportional fall

After the years of scorched earth budgeting under Connex, the state run South Eastern Trains is now starting to spend money on the simple things – one of which is appointing a Traction & Rolling stock Engineering Director. Apart from the fact that the job title is wrong – it should be Chief Mechanical & Electrical Engineer of cource, this is timely move – and his first question could be what has happened to Class 365 reliability.

Table 2

Three phase drive EMU reliability – Period 10

Source: NFRIP

Coherent

Sorry, I've left Siemens choking. As you can see Arriva Trains Northern hasn't got the message and the Class 333 fleet is now fourth from the bottom, rather than third last time, and this is only because the Class 360s are now included.

How can this be? Well, here are the background data from the bottom of the three phase drive EMU table.

Table 3

Three phase drive EMUs

Casualty data Period 10

Source: NFRIP

As you can see from Table 1, the contractual record shows 3-5 casualties per month, while NFRIP data for the most recent four week period shows 20 – say five times.

Now look at the bottom three. Very similar mileages and an almost identical number of casualties in the four weeks from three different trains from different TOCs. And I have mentioned already the claimed differences between Contractual and NFRIP figures for the Class 458.

So despite the complaints in Forum, I reckon NFRIP data may be ‘noisy' and it may be ‘dirty', but is probably coherent and analyses like this are useful rather than merely diverting.

There is another interpretation of the difference between NRIP with and manufacturers contractual figures. UK MPC figures are could well be pessimistic in global terms.

For example, assume that foreign railways – which always have the national interest in mind – go for ‘contractual failures'. And then assume that the Class 333 factor of five applies across all fleets.

Then the best English Electric powered kit on SWT is doing over million miles per casualty on preliminary Period 11 figures and the WAGN Class 365 averaged just a quarter of a million miles over the last 12 months. These are the sort of numbers you get from patriotic railways.

Fallacy

It is also instructive to look at the point made by ATOC in Forum last month, namely that these comparisons are potentially fallacious because of the different performance of two Class 317 fleets on WAGN. As Table shows the ratio of fleet size to number of casualties in P10 is roughly the same, but the Class 317/7s do indeed run substantially more miles and stop less.

Table: 4

WAGN EMU reliability

But as Colin Boocock's letter in the same Forum pointed out, attention to detail during maintenance makes all the difference to reliability, and having cried ‘fallacy' ATOC in its penultimate sentence changed tack and conceded that maintenance needs to match the way the assets are being used.

Picking up this point, an old BR hand, now working for a metro in the sun, reminded me that in the early 1990s the BR infrastructure locomotive fleet was given this treatment.

These locos spent most of their time idling and the rest running at low speed. This meant that equipment deteriorated in different ways compared to its intended main line use.

For example, a component maintained on a distance-travelled basis could receive attention too infrequently. Similarly, engine hours spent idling or on low load, rather that at high power, overstated the need for attention to some power system components and understated others.

Following analysis of the actual working regime for each system or component the maintenance schedule for these locomotives was rewritten to match. The result was that a fleet of cast-offs and hand-me-downs performed far better (and for far longer) than anyone had ever expected.

Bring it on

Meanwhile, two high profile classes are still missing from, NFRIP – the Class 350 Desiros on SWT and Virgin's Class 390 Pendolinos.

SWT tell me that they are getting between 14,000 and 25,000 MPC from their Desiros. But of course, Period 10 started just after I handed over the bottle of bubbly to David Wilson. Perhaps next period then?

And there are dozens of Pendolinos running around, which makes their absence from NFRIP slightly more noticeable. According to Informed Sources they got up to 8,000 MPC running without tilt. As tilt has been activated for driver training more new things to go wrong have started working, pushing the engineers back up the bathtub curve. Currently the MPC is around 4000.

But of course, what people tell you is one thing, the cold numbers, quite another.

Continues.........

Return to Archive -by date - by topic.