ERTMS problems

[John_Hughes]

Nightmare journey back from Birmingham on the Cambrian main line last night, when the lead unit's ERTMS unit failed just east of Newtown. We finally dragged into the Down platform, then sat there while crew and Control discussed what to do next.

 

The decision reached was to run the lead unit round to the back, decanting all the passengers into the second unit, which ended up packed like a sardine can. The run-round operation itself took an astonishing hour to complete while the Up train was stuck at Caersws. We finally left about 90 minutes late, with the whole train taken out of service at Machynlleth with the passengers being bussed to onward destinations - and it's a long bus trip up the coast to Porthmadog!.

 

It seems that this has not been an isolated problem - the ERTMS units themselves have what seems to be anectodally speaking an unacceptably high failure rate, and when that happens even routine manoeuvres like a run-round (five minutes in the old days!) take for ever.

 

No complaints about the crew, by the way, who were doing their best as cheerfully as possible. But comments about ERTMS please? Am I being unfair?

 

[The Stationmaster]

The only thing I would say about Cambrian ERTMS is that it is a trial site and the on-train kit was a retrofit to existing vehicles (never a good way of doing such a job of course and a source of complaints since Day 1).  One can but hope that things are being developed as part of the trial and not just being ignored 'because it is only a trial site'.

['CHARD]

I have a very thick folder on my desk about this very subject that I keep putting off studying. 

 

I'm supposed to know all about the background to and intentions for the wretched system by the time of a meeting next Wednesday, so I'll see if there's anything of relevance for posting here....

[black and decker boy]

The only thing I would say about Cambrian ERTMS is that it is a trial site and the on-train kit was a retrofit to existing vehicles (never a good way of doing such a job of course and a source of complaints since Day 1).  One can but hope that things are being developed as part of the trial and not just being ignored 'because it is only a trial site'.

 

We will always be retrofitting to some degree. Whilst new units like Electrostars may claim to be compatible, until you have the exact spec of what you bolting in, it may never be a perfect match and as the trains software gets older and outdated and newer and fancier ERTMS kit comes along, these gremlins will need ironing out. That said, A 25 year old 158 must be a less user friendly interface that a 5 years old EMU.

 

Who wants to fit ERTMS to the steam railmotor then...

[corneliuslundie]

I don't understand why the unit with failed ERTMS, once at the rear of the train, could not have carried passengers. Was someone afraid it would take off on its own?

 

But I do hope the service gets a bit more reliable before I move to Newtown in the autumn. Last year we had two bad delays out of about six trips, though to be fair one was a unit delayed by a passenger taken ill somewhere Birmingham and the other was a signalling failure in the Wolverhanmpton area.

[Shed]

let's hope the trial on the Hertford loop is a success and helps to sort this kind of thing out.

[Guest Phil]

I don't understand why the unit with failed ERTMS, once at the rear of the train, could not have carried passengers. Was someone afraid it would take off on its own?

 

But I do hope the service gets a bit more reliable before I move to Newtown in the autumn. Last year we had two bad delays out of about six trips, though to be fair one was a unit delayed by a passenger taken ill somewhere Birmingham and the other was a signalling failure in the Wolverhampton area.

 

 

Must admit that crossed my mind too, but I guess it is something ridiculous like Newtown is not signalled for Permissive working - even though ERTMS has replaced RETB, which replaced conventional signalling here. FWIW a train cannot be signalled into the platform if another train is already standing there, unless it is approved for permissive working. I'm guessing that if the lead unit was split, and drawn forward onto the single line, run back through the Up platform, then driven back into the Down platform, it would be far too risky a manouvre with passengers on, so they would have to be tipped out onto a freezing cold platform for the duration of - what did you say ? An hour for gods sake !!!!

 

For those who don't know, the Cambrian network returned to the GW zone of Network Rail controlled from Swindon. The servers for the ERTMS system are located at Didcot, and Ansaldo who are contracted to maintain the system are at Stoke on Trent. Nothing like a Welsh railway run by "Jac y Saes" * is there ?

 

 

* You need to understand Welsh for that !!!

Edited February 24, 2013 by Phil

[Guest Phil]

 

A few things:

• Ansaldo have some major problems on the Continent, with both the Danish and Dutch. But also Italian Justice for corruption. It could be the final nail in the coffin, so other projects may not have that much priority right now...
• Although ERMTS is an EU spec, implementation is left to individual manufacturers. They make sure their system is only compatible with their own stuff, any competitor should adhere to their standard, but they then plainly refuse to tell what that standard then is, for competitive reasons...
• If the ERMTS installation isn't programmed to perform certain trivial tasks, it'll take the operators some time to find out and find a workaround. This workaround will take more time to perform.
• Drivers are weary to simply switch off the onboard ETCS unit, afraid that they'll be told off by their superiors for doing so and/or causing the total disabling of their train, from which they can't recover.
• ETCS units are basically computers with associated boot-times. Changing cab will power up the new unit from cold, and the driver must then put in all relevant data about his train, which takes considerable time, especially if said ETCS units are also programmed to perform a full system check on the onboard safety systems (breaking, etc)...

Maybe so, but if there are only two tracks in Newtown station, and the points are "train operated trailers" - what on earth can go wrong - that takes over an hour to shunt ?

The driver simply splits the units, drives the leading unit over the sprung points at the Caersws end of the station, changes ends and drives through the up platform, over the sprung points, changes ends, and drives this unit onto the rear of the other set. Irrespective of whether there is permission or not, the movement can be done at safe walking pace, and there are no points and signals to worry about becuase it is all automatic.

 

All that has happened is another trainload of very disgruntled passengers were made to stand on a freezing cold platform whilst the railway staff battled with the most simple of problems, whilst tied up in red tape. It seems to me that ERTMS is simply not robust enough to cope with 24 DMUs and about 150 miles of branchline railway.

[John_Hughes]

Right - from the responses so far l take it that it indeed isn't fit for purpose!

 

In fairness to the staff I must point out that no-one was tipped out to wait on the platform - everyone was shepherded into the 'rear' unit, so no cases of frostbite to add to the general fuss - and there was a general distribution of free tea and coffee as well! As I said before, all credit to the folks involved who really did do their best.

 

Regarding ERTMS reprogramming when ends are changed, at Salop this is presumably done on a regular basis as the Cambrian line trains all reverse there; doesn't seem to take more than a few moments then! And as for the empty unit running round 'at a safe walking speed', I'd love to be able to walk that fast!

[Edwin_m]

Maybe so, but if there are only two tracks in Newtown station, and the points are "train operated trailers" - what on earth can go wrong - that takes over an hour to shunt ?

The driver simply splits the units, drives the leading unit over the sprung points at the Caersws end of the station, changes ends and drives through the up platform, over the sprung points, changes ends, and drives this unit onto the rear of the other set. Irrespective of whether there is permission or not, the movement can be done at safe walking pace, and there are no points and signals to worry about becuase it is all automatic.

 

 

Didn't they take out the "sprung" points as part of the ERTMS work?  If so then presumably the move suggested would have to be "signalled" and therefore may not possible with ERTMS out of use on the unit in question? 

 

Even if the points remain sprung, wouldn't this shunt involve a move into the section?  If so authority would be required, which again would need would need a working ERTMS and wouldn't be granted if another train already had a conflicting movement authority.

[Rob Rossington]

even with sprung points the move would have to be "signalled" albeit through the use of stop boards and point indicators etc.

 

You would think that there is a contingency plan to deal with running a unit round with a broken ERTMS module. surely its possible for Machynlleth to give the driver permission to proceed past the block marker?

[Oldddudders]

even with sprung points the move would have to be "signalled" albeit through the use of stop boards and point indicators etc.

 

You would think that there is a contingency plan to deal with running a unit round with a broken ERTMS module. surely its possible for Machynlleth to give the driver permission to proceed past the block marker?

There certainly needs to be such - but new things don't always get all the necessary support documentation on Day 1, or, if they do, it may need to be revised in the light of experience. If Machynlleth does not have authority to do so, that's that.

 

In 1981, a colleague was telling me about the time he had been on a test trip with the prototype 4-PEP units. A problem had arisen and he had authorised the units to be run round to overcome this. On return to Waterloo, someone in a suit was dancing up and down, pointing out to my colleague that the special instructions specifically forbade running round the units. "I know" he replied "I wrote that instruction!"

[Eggesford box]

I am not familiar with the system so just a thought but if the unit had to go into the section to run around and the ERTMS was not working would you have to put in some form of pilotworking? May explain why it took an hour to make the move, appoint a pilotman who had to get out to Newtown.

[John_Hughes]

I'd wondered about pilotman working - in the old days if there had been a complete failure of the system it would have been instituted without (too much) delay, and he (always he!) would have been on-site within perhaps thirty minutes or less once Control had approved it. But no pilotman was in evidence, and it seems that fiddling about at Mach was the cause of the delay; once the run-round was approved the move took place reasonably quickly.

 

When I consider how much care was taken with earlier systems to ensure that a reasonable service could be maintained when things went wrong, it seems rather worrying that a minor electronic glitch can cause such chaos, with serious delays to passengers no doubt propagating through the system, to say nothing of units and possibly staff in the wrong place.

[The Stationmaster]

I'd wondered about pilotman working - in the old days if there had been a complete failure of the system it would have been instituted without (too much) delay, and he (always he!) would have been on-site within perhaps thirty minutes or less once Control had approved it. But no pilotman was in evidence, and it seems that fiddling about at Mach was the cause of the delay; once the run-round was approved the move took place reasonably quickly.

 

When I consider how much care was taken with earlier systems to ensure that a reasonable service could be maintained when things went wrong, it seems rather worrying that a minor electronic glitch can cause such chaos, with serious delays to passengers no doubt propagating through the system, to say nothing of units and possibly staff in the wrong place.

Instituting Pilot Working nowadays is in some respects simpler but I don't know about the situation on ERTMS lines (a former collegue of mine wrote the Regulations and they were based on contemporary Instructions for other systems so Pilot Working per se shouldn't be too difficult I would have thought).  It sounds in this case as if the problem was largely a technical one leading to it not being possible to find the wood among the trees in the ensuing dscussion - and that is often how the most time is lost.  Obviously - sitting here at my computer - the logical thin g to do would be to put the rear unit, with working ERTMS, onto the front of the failure; assuming no other trains were about it could readily be given proper electronic authority to set back into the section in rear and to enter tehs ection in front.  Equally obviously this wouldn't be possible if that loop still has self-acting points but (like someone earlier) I though they had been converted to power operation as part of this scheme (but do they have 'funny' control constraints I wonder?).

 

Simply put - we don't know enough to know why who decided/did what for which reasons but it does have more than a hint of boing a tad mishandled, if only in the amount of time it took.

[Talltim]

The simplest thing would be to design ERTMS so that as long as a working unit is present in a train it doesn't matter where it is

[The Stationmaster]

The simplest thing would be to design ERTMS so that as long as a working unit is present in a train it doesn't matter where it is

Alas that won't work as a principle because the train might no necessarily stop in the correct place when the ATP function requires it to do so.  I believe there were some problems because of this in the Cambrian design and it was one of the reasons why an early idea of using an ERTMS fitted diesel loco inside a steam loco was rejected.  

[ess1uk]

let's hope the trial on the Hertford loop is a success and helps to sort this kind of thing out.

I think it's 2 different manufacturers so may not help the Cambrian kit?

[ess1uk]

On 24/02/2013 at 15:28, Shed said:

let's hope the trial on the Hertford loop is a success and helps to sort this kind of thing out.

this is to test Degraded Mode Working System. the idea is to independently prove a train's position and that the track ahead is safe to proceed, then to issue a movement instruction to the driver via GSMR

[corneliuslundie]

I had forgotten about this discussion.

Just to say that a couple of years ago we were on a train from Shrewsbury to Newtown which kept coming to as halt in section. It was concluded that the connections between the two units had become dirty and electrical contact was intermittent - every time it was lost it caused the brakes to go on. The driver explained this over the intercom.

When we got to Welshpool they did exactly as discussed. The units were separated, the front unit ran  forward, back through the up platform and then onto the rear of the train. It took just a few minutes with no-one detrained. Problem with the brakes solved. Evidently some lessons had been learned since 2913.

Jonathan