Easter engineering works and subsequent ongoing disruption - April / May 2015

[whole diary]

[BerkshireBugsy]

Rather better than the Chiltern line today (unusually) which has seen trains over two hours late and multiple cancellations from 8:30am this morning following problems at Wembley and Banbury.  Good job it wasn't the play-off final today!

Very good point. I imagine the worst case scenario would be if spectators were delayed by 30 minutes getting in, their team loosing the match and then being delayed 2 hours getting home...

[ChrisB]

Which play-off final?...the one on May 25th will be affected by the NR^» (Network Rail - home page) strike!

[TaplowGreen]

Which play-off final?...the one on May 25th will be affected by the NR^» (Network Rail - home page) strike!

......and the one on the 24th - Swindon have got to Wembley - but no trains from Swindon-Paddington.

[stuving]

I'm led to believe that the new axle counters are struggling to always give the correct readings due to quality of the track - especially in the West Drayton area - and then fail safe.  I'm sure that when they bed down then matters will improve significantly, but the on-going disruption is disappointing.

I'm still puzzling over how that can happen. After all, much higher reliability is generally seen as one of the two main advantages of axle counters over track circuits. (The other, which prompted their use in the current Western Route deployment, is their inherent immunity to traction currents.)

Put another way, any deficiency in the track that can cause a miscount has to either generate a false count or suppress a true one, or so distort the expected signals as to raise an error flag. That third possibility sounds most likely, but even then surely it would take a pretty gross track problem to do it? Or maybe it's an installation issue - such as a detection head being loose - in which case it will not "bed down" without help. So what kind of poor quality track is involved here?

As an aside, track circuits can detect track faults and that is claimed as an advantage, since axle counters can't. In reality a track circuit can only reliably detect a full break in a rail, and you need to be finding the cracks well before that on this railway. A barely-maintained freight-only line in the wilds of the USA is different.

I ought perhaps to declare that I used to work for Thales, but never had anything to do with railway signalling. I have since heard that Thales UK^▸ (United Kingdom) does now do some design support engineering work on axle counters, perhaps because there are so many of them being installed. But I can claim to not be an expert on the subject.

[TaplowGreen]

The latest but there have been others between Reading and Paddington today - despite all the ongoing work reliability seems to be getting worse rather than better.

Due to signalling problems at Maidenhead trains have to run at reduced speed on some lines.
Impact:
Train services running through this station may be delayed by up to 15 mins. Disruption is expected until 20:00 18/05.

[a-driver]

I'm led to believe that the new axle counters are struggling to always give the correct readings due to quality of the track - especially in the West Drayton area - and then fail safe.  I'm sure that when they bed down then matters will improve significantly, but the on-going disruption is disappointing.

I'm still puzzling over how that can happen. After all, much higher reliability is generally seen as one of the two main advantages of axle counters over track circuits. (The other, which prompted their use in the current Western Route deployment, is their inherent immunity to traction currents.)

Put another way, any deficiency in the track that can cause a miscount has to either generate a false count or suppress a true one, or so distort the expected signals as to raise an error flag. That third possibility sounds most likely, but even then surely it would take a pretty gross track problem to do it? Or maybe it's an installation issue - such as a detection head being loose - in which case it will not "bed down" without help. So what kind of poor quality track is involved here?

I don't think it's related to a deficiency in the track either.  When the axle counters fail the signaller tries to perform a reset on the system, this doesn't work.  If it was a track related fault then I would have thought a reset would work??

[stuving]

I don't think it's related to a deficiency in the track either.  When the axle counters fail the signaller tries to perform a reset on the system, this doesn't work.  If it was a track related fault then I would have thought a reset would work??

Once the count is wrong, it has to be reset. Since that can make an occupied block show as clear, it's a very controlled process. One way is to set it as "unsafe" and then talk one train through, so the count goes up and then down to zero again. Then it shows "clear". Of course this takes some time.

If the counting process is not working properly, for whatever reason, you can go through all this process and end up with it still showing occupied when it shouldn't. Of course the counter (or ACE - Axle Counter Evaluator) is meant to detect most errors itself, in addition to their being very rare.

[TaplowGreen]

So why do they keep failing?

[SandTEngineer]

I understand that there is a technical problem with the Frausher detector heads themselves.  Apparently there is a defect which is revealed when there is a very rapid external temperature change.  This problem has also manifested itself on the Cardiff resignalling area (although the failures there don't get as much publicity as elsewhere).

[stuving]

I understand that there is a technical problem with the Frausher detector heads themselves.  Apparently there is a defect which is revealed when there is a very rapid external temperature change.  This problem has also manifested itself on the Cardiff resignalling area (although the failures there don't get as much publicity as elsewhere).

Curiouser and curiouser. Frauscher are an Austrian company, so it's hard to believe the temperature range in Cardiff is outwith their experience. And the spate of failures has been on "Crossrail West Outer" - Airport Junction to Maidenhead - where I though NR^» (Network Rail - home page) were putting Thales AzLMs.

I found this press release, dated 8th April, with two interesting pictures. The first isn't Frauscher kit - they don't so yellow mushrooms. The second is Slough, and if you look back on this thread to what was happening there on April 7th and 8th (and since), you do wonder whether Technotrack's message wasn't both premature and too self-satisfied.

[phile]

Rather than discussing Engineering Train Alterations, items which are better suited to the "Thames Valley Signalling Problems etc" Thread seem to have crept in here..