Serious Accident in Bavaria

[EAD]

Yes editing is an issue with such statements! The BBC TV news at lunchtime had a reporter in the studio essentially tell viewers single track lines are inherently dangerous. Objectively in theory yes, but of course no explanation added by him that this is why signal interlocking, etc. exists.

Quite heated debate going on over at our German equivalent Drehscheibe-Online. All speculation, but put simply in normal operation this crash cannot happen so we need to wait and see what was going on to allow a second train into the block.

I should add there is a splitting block signal at the halt Bad Aibling Kurpark, but this is merely there to split the length of the block between the stations and also protects a level crossing (i.e. does not clear until the barriers are secured), but there will only be one single route that can be set for one direction or the other between Bad Aibling and Kolbermoor.

 

[Merseysider]

Human error was probably to blame, according to the initial investigation, reported by der Spiegel (in German).

The person responsible hasn't been named but there doesn't appear to be a mechanical fault involved. Latest figures are 10 dead, 18 serious injuries and another 90 minor casualties with one person still missing. Hopefully that won't increase any further.

 

[amcluesent]

Suggestion that it had become the habit of staff controlling the signals to disable the safety system to wave through a late-running train and today they overlooked the other train, in echoes of Quintinshill.

 

[w0033944]

Suggestion that it had become the habit of staff controlling the signals to disable the safety system to wave through a late-running train and today they overlooked the other train, in echoes of Quintinshill.

If that dos, in due course, prove to be the case, I suspect the line operators will be in hot water. Such flouting of fundamental procedures is sure unforgivable.

 

[EAD]

There had been some pretty poor reporting tonight in Germany (those who think that is a UK only issue take note). Spiegel and others quoting sources pointing finger at the signalman, but it is far too early to know frankly even though of course those on the ground will have a very quick idea of what was going on signalling wise.

I think there is a lot of speculation based on not understanding how PZB works etc. There is no magic way for a signalman to override or switch off PZB as some media is suggesting. What there is is a procedure for authority to pass a signal at danger - it is different to here and involves the driver receiving the authority and holding down a button to override the PZB brake application - they are then restricted to 40Km/h until clear of the points area or they can see the next signal and can return to line speed.

Here the odd thing is how or why would the signalman have given that when he is in charge of both trains and would hardly be overcome by this standard situation which happens once an hour or more. Even if the signal had a fault, there are procedures to follow around other trains before you get to setting up a route and giving an order - the interlocking would still be there to try and stop you doing it if a route is already set.

 

[NicholasNCE]

Thanks EAD for sharing your knowledge of the PZB protection sytem, a great insight!

Here's a cab-view video of the line: https://youtu.be/y04O09B8vy0?t=25m22s (link should begin at the entry of the concerned single line section)

 

[zuriblue]

While it's been suggested the guard may have been in the cab, on a quiet regional service most people will be sitting down in what are mostly (on those trains) airline seats. If the guard was standing in the aisle checking tickets, he could well have been thrown the full length of the train.

Especially since FLIRTS are articulated units with Jacobs Bogies so there isn't much to stop someone getting thrown the length of the train. It doesn't help that the units are quite powerful so accelerate quickly (a 4 car FLIRT weights around 120 tones and develops around 2000 KW at the wheel - in comparison a Class 442 is around 200 tones and develops around 1200 KW)

 

[30907]

Anyone who has good German will find the post by Ex-bahner on p23 of this thread informative. http://www.drehscheibe-online.de/foren/read.php?2,7720582,page=23. Google translate doesn't do a good job!
As EAD says, the rest of the thread is pretty speculative, and the media even more so.

 

[D1009]

Thanks EAD for sharing your knowledge of the PZB protection sytem, a great insight!

Here's a cab-view video of the line: https://youtu.be/y04O09B8vy0?t=25m22s (link should begin at the entry of the concerned single line section)

Thanks for posting that link, I presume the entry point is Bad Aibling?

 

[BRX]

My attempt at a translation of the last bit of that post from the German forum -

The dispatcher could have postponed the crossing-over of the trains until Bad Aibling, so that the train traveling towards Holzkirchen train didn't build up too much delay in Kolbermoor. Let's think about that. He readies the train in Kolbermoor to depart. The signal is green. Then he decides, for whatever reason, to let the train travelling towards Rosenheim leave Bad Aibling after all. He can't turn the departure signal green, because the track is occupied. He must use ZS1 ("substitute sugnal"?) / ZS7 ("caution signal"?) or command. But as soon as the train leaves Bad Aibling station and occupies the section, the signal in Kolbermoor must turn to red. If the train has not yet passed it. The distant signal for signal 314 is located only a few hundred meters away from the accident scene, perhaps the driver had even already seen the yellow. Because even with the crossing barriers closed, it would not have been allowed to show permission any more, if the train coming from Bad Aibling train was already in the section.

Let us go back to the train travelling towards Rosenheim: The train therefore departs, then stops at the Kurpark halt. And is back to a red light at signal 313. Here, the situation is the same: Even this signal will not turn green, because the section is occupied towards Kolbermoor. Even if the barriers are closed. The dispatcher would have to again use ZS1/ZS7 over-rides. And close the level crossing manually. And in spite of that, forget that he had indeed allowed the train to leave Kolbermoor. The track section would be illuminated red.

Seems to me, rather a lot of blackouts.

Track diagram is here:

http://www.directupload.net/file/d/4259/uhnisk8n_jpg.htm

 

[kevconnor]

The latest report from the BBC is reporting the discovery of what it is calling the trains 'black box.' Whilst I am guessing it is using this language as it is synonymous with an airplanes flight recorder, can anyone describe what it is and what it records, (I am guessing telemetry). Are these, or something similar, found on British trains?

 

[EAD]

Anyone who has good German will find the post by Ex-bahner on p23 of this thread informative. http://www.drehscheibe-online.de/foren/read.php?2,7720582,page=23. Google translate doesn't do a good job!
As EAD says, the rest of the thread is pretty speculative, and the media even more so.

Yes - Ex-Bahner has quite a good summing up there and some comments on some of the more wild speculation.

Thanks for the nice comments generally - as someone interested in German railways as well as ours nice to relay things, though I wish it was under better circumstances.

D1009: yes that video starts at Bad Aibling heading towards Kolbermoor. The impact was just into the left curve next to the sewerage works (l/h side) and the Manfal canal (r/h side).

 

[Bletchleyite]

<delete>

 

[sng7]

The latest report from the BBC is reporting the discovery of what it is calling the trains 'black box.' Whilst I am guessing it is using this language as it is synonymous with an airplanes flight recorder, can anyone describe what it is and what it records, (I am guessing telemetry). Are these, or something similar, found on British trains?

Yes the uk has On train Monitoring Recorders (OTMR) fitted to rail vehicles on the national network which record information like and aircraft black box would.

 

[Bletchleyite]

This is a link to a translation of the description of the Zs1 (and you can link from there to the Zs7) "override signal". It appears this is a different way of instructing a driver to pass a signal at danger then drive on sight as would be done in a more manual way in the UK:

https://translate.google.com/transl...&u=https://de.wikipedia.org/wiki/Ersatzsignal

However...would that be permitted onto a single line in the UK? It seems a very dangerous concept that it even exists as an option on a single or bidirectional line indeed.

If a signal fails on a single line in the UK, what is the procedure to pass it at danger? Must the line be checked and a pilotman appointed?

 

[ComUtoR]

This is a link to a translation of the description of the Zs1 (and you can link from there to the Zs7) "override signal". It appears this is a different way of instructing a driver to pass a signal at danger then drive on sight as would be done in a more manual way in the UK:

It sounds like a PoSA




Do they have FFCCTV (forward facing CCTV) on their units ?

 

[Merseysider]

First of all, here are the relevant sections of the timetable, shown as Station - Arrival - Departure - Delay.

Bad Aibling 0637/0638 +4
BA Kurpark 0640/0640 +4
Kolbermoor 0644/0644 +4
->Rosenheim

Kolbermoor 0640/0645
BA Kurpark 0648/0648
Bad Aibling 0650/0651
->Holzkirchen

And now, here's the post on Drehscheibe.

"I offer a summary of the facts, based on my own knowledge and without speculation. There are countless posts on this thread and numerous reports currently circulating in the media containing fiction so I'll restrict what I mention to that which is conclusive based on the evidence and verifiable by fact. Anything else (ie allegations, speculation, suggestions) will be ignored. This is an attempt to dispel the (tasteless) rumours which I view with much disdain.

Both Bad Aibling and Kolbermoor stations are controlled by the signalbox at Bad Aibling. Whether the signalbox operations are done there, or from the Operations Centre in Munich, makes no difference. Between BA and KO stations there is a smaller halt by the name of Bad Aibling Kurpark, which has Automatic Signal Block #313 at the end of the platform (towards Rosenheim). In the other direction (towards Holzkirchen) lies Automatic Signal Block #314. Both of these determine whether the half barriers (remotely operated) at the Rosenheimer Street level crossing are up or down.

The maximum line speed in this area is 100km/h (60mph) before increasing further down the line to 120km/h (70mph).

Both trains should have arrived into Kolbermoor (where the single line becomes double line due to the platforms) at roughly the same time; at 0644 and 0645 hours. According to the Police Department of Upper Bavaria (South) at 0648 the first emergency call was made. The 0644 train to Rosenheim was running 4 minutes late.

These are facts.

This evening, it was reported in the media that, according to the investigation team, both trains had been erroneously dispatched to the same section of track by the signaller. Police have rejected the fact that the investigation teams were disemminating information.

I have firsthand knowledge of the locations involved.

The signaller could have withheld the Rosenheim-bound train from the junction to prevent the (punctual) Holzkirchen train from being delayed too. However, the train sitting in Kolbermoor (the 0645 departure to Holzkirchen) was given the green light. Then, for whatever reason, the signaller decides to allow the delayed train sitting in Bad Aibling Kurpark (which, if it had been punctual, would have been in Kolbermoor at this point) to proceed towards Rosenheim at 0646.

Neither train would have received any notification of the oncoming train; for all the driver knows the track is closed to all other vehicles and only he can access the single line section of track. As soon as the train leaves Bad Aibling station, the signal in Kolbermoor is automatically set to 'STOP' and vice versa.

Automatic signal block #314 lies only a few hundred metres from the scene of the crash and it stands to reason that the train driver (of the Holzkirchen-bound train) would have seen the negative aspect (red signal) too late to do anything.

With the level crossing barriers down, and the Rosenheim-bound train already in the section, the Holzkirchen-bound train should have received no signal whatsoever to proceed. This is a point the investigation will be focusing on.

Going back to the Rosenheim-bound train (running 4 minutes late); it departs Bad Aibling at 0642 and shortly afterwards pulls into Bad Aibling Kurpark. And here, the situation is the same; Automatic Signal Block #313 shows a stop signal due to the presence of a train in the section between there and Kolbermoor.

At 0646, the signaller would have manually reset Substitute Signals 1 and 7, and then immediately afterwards operated the level crossing barriers to prevent cars crossing the track so that the Bad Aibling-Kolbermoor train can proceed. Then - inexplicably - he has given the Kolbermoor - Bad Aibling the signal to proceed, at 0646. Even though the light is showing red. That's one hell of a lot of blackouts. (Draw your own conclusions). But like I said earlier, I'd rather stick to only facts."

I've attached a sketch provided by the poster in the German forum.

 

[BRX]

^ I don't think that is an entirely correct translation of what the poster on the German boards actually wrote.

 

[Merseysider]

^ I don't think that is an entirely correct translation of what the poster on the German boards actually wrote.

It's not word for word. If you translate word for word you end up sounding like a robot or missing the meaning. For example, Fdl, the German abbreviation for Fahrdienstleiter, doesn't mean dispatcher in this context. It refers to the signalling box, ie one of these.

 

[Jonny]

It's not word for word. If you translate word for word you end up sounding like a robot or missing the meaning. For example, Fdl, the German abbreviation for Fahrdienstleiter, doesn't mean dispatcher in this context. It refers to the signalling box, ie one of these.

This one would be consistent with US English, where Dispatcher is equivalent to Signalman in British English.

 

[Groningen]

Question remains whether both trains had a green signal. It also happened in a bend; although it was still dark the headlights of both trains must be viseble to both drivers if it was a straight line of tracks.

 

[BRX]

It's not word for word. If you translate word for word you end up sounding like a robot or missing the meaning. For example, Fdl, the German abbreviation for Fahrdienstleiter, doesn't mean dispatcher in this context. It refers to the signalling box, ie one of these.

Ok, but for example he says:

Das Vorsignal des Skb 314 dürfte sich nur weniger hundert Meter von der Unfallstelle entfernt befinden, vielleicht hat der Tf sogar noch das Vr0 gesehen.

Which you've translated as

Automatic signal block #314 lies only a few hundred metres from the scene of the crash and it stands to reason that the train driver (of the Holzkirchen-bound train) would have seen the negative aspect (red signal) too late to do anything.

But I think he's actually saying that the *distant* signal for 314 is a few hundred metres from the scene of the crash, and that the driver *might* have already seen it at a *yellow* aspect before the crash.

 

[Bletchleyite]

It sounds like a PoSA

A what, sorry?

 

[GB]

Proceed on Sight Authority.

If there is a problem with the signalling, instead of the usual method of bringing the train to a stand, communicating with the signaller and then moving again, the driver can just keep going at caution if the PoSA is flashing. Saves time and thus saves money.

 

[Bletchleyite]

Proceed on Sight Authority.

If there is a problem with the signalling, instead of the usual method of bringing the train to a stand, communicating with the signaller and then moving again, the driver can just keep going at caution if the PoSA is flashing. Saves time and thus saves money.

Ah yes.

The problem on a single line, of course, is that the *other* train is proceeding *not* on sight.

That approach only works safely on a unidirectional line, where the worst case is a stationary train, not one proceeding towards you at 100km/h.

 

[MarkyT]

It sounds like a PoSA

A what, sorry?

POSA = "Proceed On Sight Aspect"

A relative newcomer to UK, consisting of an extra position light aspect on a plain line main signal for the signaller's use in authorising movements quickly under certain failure conditions without having to communicate verbally with the driver. In Germany they have had something similar for decades and they are known as a 'ersatzsignal' which translates to 'substitute signal'. They shouldn't be able to defeat opposing locking controls though, only things like track circuit failure or lamp out ahead, possibly loss of trailing point detection. Germany has had an incident with these before, the Bruehl dertailment in 2000 was a misinterpreted ersatzsignal, used for wrong line running during engineering works in that case:
https://en.wikipedia.org/wiki/Brühl_train_disaster

 

[EAD]

Ah yes.

The problem on a single line, of course, is that the *other* train is proceeding *not* on sight.

That approach only works safely on a unidirectional line, where the worst case is a stationary train, not one proceeding towards you at 100km/h.

Yes - the thing is that in Germany the onus is on the signaller where there is a fault and there are procedures so any form of Zs signal cannot be used without clear procedures to check the status of the route set and confirm it is locked. It will be very interesting to hear exactly what happened here as somewhere the rules and procedures at first sight have not been followed. I find it very hard to fathom how, given the same signaller was in charge of both trains on his panel, he could have hypothetically convinced himself there was not a train in the section and go through all the steps to give authority to enter even though it would clearly show as occupied/a normal route would not release and that would be contrary to the rules and common sense. We will have to wait as it does not add up based on what we know/how the rules and systems work.

 

[Bletchleyite]

Yes - the thing is that in Germany the onus is on the signaller where there is a fault and there are procedures so any form of Zs signal cannot be used without clear procedures to check the status of the route set and confirm it is locked.

But you could equally say interlocking was unnecessary because there are procedures.

Unless those procedures involve multiple safeguards - such as the involvement of multiple people - this seems a significant safety flaw.

 

[matchmaker]

But you could equally say interlocking was unnecessary because there are procedures.

Unless those procedures involve multiple safeguards - such as the involvement of multiple people - this seems a significant safety flaw.

Edward Tyer must be turning in his grave! Since Abermule in 1921, as far as I know, there has only been one single track head on collision in the UK caused by human error.

 

[Bertiethebobby]

Yes - the thing is that in Germany the onus is on the signaller where there is a fault and there are procedures so any form of Zs signal cannot be used without clear procedures to check the status of the route set and confirm it is locked. It will be very interesting to hear exactly what happened here as somewhere the rules and procedures at first sight have not been followed. I find it very hard to fathom how, given the same signaller was in charge of both trains on his panel, he could have hypothetically convinced himself there was not a train in the section and go through all the steps to give authority to enter even though it would clearly show as occupied/a normal route would not release and that would be contrary to the rules and common sense. We will have to wait as it does not add up based on what we know/how the rules and systems work.

I agree it does seem hard to understand how a signalman who controls both ends of the section could do that. Perhaps that is the key though. On most single track sections the signalling is done by co-operation of the 2 signalmen and one would hope that if went to do something unsafe that the other would pick up on it and question it. Inexplicable things do happen, which seem ridiculous in the aftermath. Unfortunately there are many instances of signalmen convincing themselves of a certain situation, only for it to be proved catastrophically wrong. 99.9% of the time the system will stop the signalman making a fatal error, but not always.
I very much hope as one myself that it wasn't the signalman's fault on this occasion.
There but for the grace of God go all of us.
--- old post above --- --- new post below ---

Edward Tyer must be turning in his grave! Since Abermule in 1921, as far as I know, there has only been one single track head on collision in the UK caused by human error.

There was a minor collision at Westbury on the Cambrian line in the 80's The down train entered the loop too fast, couldn't stop in time and went out the other end of the loop onto the single line colliding with the up train which was waiting at the Home signal.