Mill Hill again.

[GentlemanJim]

Looks like the same thing has happened as before, surely this can't be right?

I'm sure the Evening No standards are going to have a field day

[Nortube]

and no doubt the rmt will be jumping up and down as well.

As long as there was only one train on the single line in both incidents then, under old terms, in can be thought of as having speed control signalling between the platform and the points to the single line, with various speed signals between the starter and the points. A train is allowed to leave the platform and proceed as far as the points, the speed signals ensuring the train’s speed is reduced sufficiently to maintain the “safe braking distance” between the train and the points.

What the Northern line should do is just put out a briefing to the effect that the new signalling system allows such moves. The new system is a big change to what people have had engrained in them over the years regarding what can and can’t be done. It is much simpler to grasp for a new driver who knows nothing about the workings of the old signalling system.

[drico]

It seems to have been the controllers using click and drag, not the fault of the TBTC system.

[peteuxb]

Oo-er, this is a completely different world to the one I was familiar with on LU! There was no "click and drag" in a lever frame, or on a programme machine (even when in 'button'), so perhaps a new form of locking needs to be introduced to prevent set-up of conflicting moves to/from Mill Hill East.
(??)

[GentlemanJim]

Aug 9, 2013 10:36:11 GMT drico said:

It seems to have been the controllers using click and drag, not the fault of the TBTC system.

Can you explain 'click and drag' in signalling terms as it's not something I ever came across at the Central SCC.

[railtechnician]

Aug 12, 2013 2:57:05 GMT peteuxb said:

Oo-er, this is a completely different world to the one I was familiar with on LU! There was no "click and drag" in a lever frame, or on a programme machine (even when in 'button'), so perhaps a new form of locking needs to be introduced to prevent set-up of conflicting moves to/from Mill Hill East.
(??)

There should be non-safety interlocking in the controls to prevent conflicting routes being set! IMHO there was absolutely nothing wrong with standard traditional LT signalling with mechanical, electrical, electronic and computer interlocking. If the non-safety interlocking failed or was otherwise circumvented by fault or failure the safety interlocking prevented disaster. Sadly the trend towards removal of lever frames to be replaced by PLCs may perhaps mean that without very careful scrutiny and testing of new interlockings the railway is less safe than it used to be. I know nothing at all about TBTC but I would've thought software interlocking to prevent conflicting moves being setup or to guarantee the safety of preselected conflicting moves was absolutely fundamental in any system where lever frame mechanical safety interlocking has been removed.

It matters not what methods of signal control are employed, the last line of defence so to speak is what is at trackside. Of course I don't know the exact nature of the incident, I don't think anyone has spelt out what occurred so I cannot say more.

[peteuxb]

My thanks to RT for his thoughts on this and, although I don't have extensive knowledge of signalling, he has my agreement. Perhaps I didn't word my post particularly well, as what I meant by "a new form of locking" was something 'new' (that is 'additional')within the present system on Northern Line to make it as robust as traditional LT signalling.

[railtechnician]

It would be nice if someone would explain what actually occurred as regards the actual train movement that resulted in an incident. The inference would appear to be that a train was sent into the single line section while it was still occupied by the previous service. However, I fail to see how the points would have been reversed to allow such a move into an occupied single line without a wrong side signal failure so was the incident a derailment or something else?

[Flybe]

What actually happened wasn't an "incident" in TBTC terms. A train was at Mill Hill and one was being held by the system quite rightly at Finchley Central until the train from Mill Hill returned.





Apparently a signaller mistook it for a High Barnet service and manually "departed" the train. This told the system to allow the train to go. Remember that this is a new signalling system, not bound by the old ways of thinking. This is not unsafe as the system would say, ok fair enough, go as far as x where you will stop at the safe distance from the train already there. It was a big inconvenience but never a safety failure.





What did not help was a call from the signaller to the driver to brake as it gave the (false) impression that they had to intervene to stop the train manually. The actual reason for the call was to try (but failed) to get the train stopped before it locked up the points which prevented the train from Mill Hill getting back to Finchley Central. A wrong direction move then allowed the train to move and the other one then to continue to Mill Hill.





Yes a big cock up from a delay to service point of view but no danger at all. Remember also that it was only due to the manual intervention to depart the train that it happened at all. This signalling system can do things old systems couldn't. An old system could only handle one train in that section, this one works on safety distances just like cars on a road except that car drivers don't have a system to automatically prevent them getting too close to the one in front! There is always a minimum distance but this can vary depending on speeds, like the M25 on a bad day cars can safely be a lot closer than when travelling at 70mph.

[railtechnician]

Sept 8, 2013 11:10:20 GMT Flybe said:

What actually happened wasn't an "incident" in TBTC terms. A train was at Mill Hill and one was being held by the system quite rightly at Finchley Central until the train from Mill Hill returned.





Apparently a signaller mistook it for a High Barnet service and manually "departed" the train. This told the system to allow the train to go. Remember that this is a new signalling system, not bound by the old ways of thinking. This is not unsafe as the system would say, ok fair enough, go as far as x where you will stop at the safe distance from the train already there. It was a big inconvenience but never a safety failure.





What did not help was a call from the signaller to the driver to brake as it gave the (false) impression that they had to intervene to stop the train manually. The actual reason for the call was to try (but failed) to get the train stopped before it locked up the points which prevented the train from Mill Hill getting back to Finchley Central. A wrong direction move then allowed the train to move and the other one then to continue to Mill Hill.





Yes a big cock up from a delay to service point of view but no danger at all. Remember also that it was only due to the manual intervention to depart the train that it happened at all. This signalling system can do things old systems couldn't. An old system could only handle one train in that section, this one works on safety distances just like cars on a road except that car drivers don't have a system to automatically prevent them getting too close to the one in front! There is always a minimum distance but this can vary depending on speeds, like the M25 on a bad day cars can safely be a lot closer than when travelling at 70mph.

Whether you regard TBTC as clever or not is rather irrelevant to the fact that it is obviously not 'intelligent'. That is not to say that it is not or that it is unsafe, but simply that the humans that created and programmed it failed to consider human error in all its forms and mitigate the consequences.
No matter how one looks at it from a purely signalling viewpoint, it should be impossible for a train approaching MHE to prevent one already there from leaving, that is the train approaching the station should never have approach locking extended to the track over the points and/or any preceeding fouling track as long as any of the tracks beyond the points and adjacacent fouling tracks are occupied either legitimately or not.
I would expect the reverse to true, i.e. that as installed no train occupying the platform would be given the starter if any tracks within the limit of control including any adjacent fouling tracks were occupied legitimately or otherwise.

Two thoughts come to mind that should be in any signalling design engineer's mind; (a) Never ASSUME anything. (b) If it is possible, it can and will happen.

Obviously even moving block signalling needs some fixed 'x' points and that of course is what appears to be missing, quite simply a train in the single track section needs to also permanently 'occupy' the last approach track before the track over the points until such time as it has reversed and cleared the single track section. Quite easy to achieve in conventional signalling although it was never done that way. The signaller can then attempt to send trains to his heart's content and not prevent a train in the single track section from leaving though he might conceivably lockup Finchley Central !!!!




 

[Deleted]

Sept 8, 2013 14:35:58 GMT railtechnician said:

Sept 8, 2013 11:10:20 GMT Flybe said:

What actually happened wasn't an "incident" in TBTC terms. A train was at Mill Hill and one was being held by the system quite rightly at Finchley Central until the train from Mill Hill returned.

Apparently a signaller mistook it for a High Barnet service and manually "departed" the train. This told the system to allow the train to go. Remember that this is a new signalling system, not bound by the old ways of thinking. This is not unsafe as the system would say, ok fair enough, go as far as x where you will stop at the safe distance from the train already there. It was a big inconvenience but never a safety failure..

Whether you regard TBTC as clever or not is rather irrelevant to the fact that it is obviously not 'intelligent'. That is not to say that it is not or that it is unsafe, but simply that the humans that created and programmed it failed to consider human error in all its forms and mitigate the consequences.

From Flybe's account it strikes me that the system is as 'intelligent' as it needs to be insofar as it was (correctly) holding the Northbound train at a point necessary to enable the preceding train to vacate the branch. The system only allowed the Northbound train to depart because its human operator told it to.

Frankly I don't see the issue of having the ability to signal a second train onto the branch. Operationally it could even be useful should it be necessary to rescue a stalled train. The assisting train could proceed onto the branch more safely and more quickly than would otherwise be necessary.

Perhaps all that needs to be considered is the provision for an "are you sure?" type message to be displayed to the signaller, requiring acknowledgement and confirmation, in the event that an attempt to manually route a second train onto the branch is made.

[railtechnician]

Sept 15, 2013 16:58:19 GMT @amershamfast said:

Sept 8, 2013 14:35:58 GMT railtechnician said:

Whether you regard TBTC as clever or not is rather irrelevant to the fact that it is obviously not 'intelligent'. That is not to say that it is not or that it is unsafe, but simply that the humans that created and programmed it failed to consider human error in all its forms and mitigate the consequences.

From Flybe's account it strikes me that the system is as 'intelligent' as it needs to be insofar as it was (correctly) holding the Northbound train at a point necessary to enable the preceding train to vacate the branch. The system only allowed the Northbound train to depart because its human operator told it to.

Frankly I don't see the issue of having the ability to signal a second train onto the branch. Operationally it could even be useful should it be necessary to rescue a stalled train. The assisting train could proceed onto the branch more safely and more quickly than would otherwise be necessary.

Perhaps all that needs to be considered is the provision for an "are you sure?" type message to be displayed to the signaller, requiring acknowledgement and confirmation, in the event that an attempt to manually route a second train onto the branch is made.

I cannot disagree more!
While it might be useful to allow a second train into the branch, that cannot be achieved with the present system otherwise there might well have been an incident of another kind! 
As for enabling an assisting train into the branch in the event of a stalled train in the branch that would be an exceptional move not routinely covered by the signalling system just as in traditional LT signalling. 

Exceptional moves take time as they have to be made ultra safely, they are not signalled but rely upon rules, regulations, procedures, knowledge, experience and trust to complete and have consequences when things go wrong for whatever reason.

As I mentioned earlier there needs to be a fixed point beyond which trains will not pass unless the branch is clear. Clearly the intention was for that point to be the starter at Finchley Central and clearer still is the fact that the starter was allowed to clear and subsequently the train proceeded to a point which fouled the other road.

The issue is apparently the innocent mistake of the signaller but the crux of the matter is that the signalling system did not prevent the train from fouling the other road. Whether the signaller is given an 'are you sure' check before clearing the starter does nothing to prevent that.

[Deleted]

Sept 16, 2013 14:54:41 GMT railtechnician said:

Sept 15, 2013 16:58:19 GMT @amershamfast said:

From Flybe's account it strikes me that the system is as 'intelligent' as it needs to be insofar as it was (correctly) holding the Northbound train at a point necessary to enable the preceding train to vacate the branch. The system only allowed the Northbound train to depart because its human operator told it to.

Frankly I don't see the issue of having the ability to signal a second train onto the branch. Operationally it could even be useful should it be necessary to rescue a stalled train. The assisting train could proceed onto the branch more safely and more quickly than would otherwise be necessary.

Perhaps all that needs to be considered is the provision for an "are you sure?" type message to be displayed to the signaller, requiring acknowledgement and confirmation, in the event that an attempt to manually route a second train onto the branch is made.

I cannot disagree more!
While it might be useful to allow a second train into the branch, that cannot be achieved with the present system otherwise there might well have been an incident of another kind! 
As for enabling an assisting train into the branch in the event of a stalled train in the branch that would be an exceptional move not routinely covered by the signalling system just as in traditional LT signalling. 

You've lost me. Plainly it can be achieved with the present system (i.e. TBTC) because that is precisely what occurred in the scenario which is being discussed here. From all the information I have been privy to on various fora from people who work with the system in one way or another, it is clear that even if another train is signalled onto an already occupied MHE branch, it will only receive movement authority up to a point a safe distance away from the train already there.

Sept 16, 2013 14:54:41 GMT railtechnician said:

As I mentioned earlier there needs to be a fixed point beyond which trains will not pass unless the branch is clear. Clearly the intention was for that point to be the starter at Finchley Central and clearer still is the fact that the starter was allowed to clear and subsequently the train proceeded to a point which fouled the other road.

Please explain why it is necessary for there to be such a fixed point, from whatever perspective you consider so important, when the system as it stands plainly will not allow two trains to come into direct contact with each other on the branch (other than ancient signalling principles which are not applicable to a TBTC railway).

Sept 16, 2013 14:54:41 GMT railtechnician said:

The issue is apparently the innocent mistake of the signaller but the crux of the matter is that the signalling system did not prevent the train from fouling the other road. Whether the signaller is given an 'are you sure' check before clearing the starter does nothing to prevent that.

What do you mean by "fouling the other road"? It was on the same road! And completely safely, albeit inconveniently as it was not intended.

[railtechnician]

Sept 16, 2013 20:04:19 GMT @amershamfast said:

Sept 16, 2013 14:54:41 GMT railtechnician said:

I cannot disagree more!
While it might be useful to allow a second train into the branch, that cannot be achieved with the present system otherwise there might well have been an incident of another kind! 
As for enabling an assisting train into the branch in the event of a stalled train in the branch that would be an exceptional move not routinely covered by the signalling system just as in traditional LT signalling. 

You've lost me. Plainly it can be achieved with the present system (i.e. TBTC) because that is precisely what occurred in the scenario which is being discussed here. From all the information I have been privy to on various fora from people who work with the system in one way or another, it is clear that even if another train is signalled onto an already occupied MHE branch, it will only receive movement authority up to a point a safe distance away from the train already there.

Sept 16, 2013 14:54:41 GMT railtechnician said:

As I mentioned earlier there needs to be a fixed point beyond which trains will not pass unless the branch is clear. Clearly the intention was for that point to be the starter at Finchley Central and clearer still is the fact that the starter was allowed to clear and subsequently the train proceeded to a point which fouled the other road.

Please explain why it is necessary for there to be such a fixed point, from whatever perspective you consider so important, when the system as it stands plainly will not allow two trains to come into direct contact with each other on the branch (other than ancient signalling principles which are not applicable to a TBTC railway).

Sept 16, 2013 14:54:41 GMT railtechnician said:

The issue is apparently the innocent mistake of the signaller but the crux of the matter is that the signalling system did not prevent the train from fouling the other road. Whether the signaller is given an 'are you sure' check before clearing the starter does nothing to prevent that.

What do you mean by "fouling the other road"? It was on the same road! And completely safely, albeit inconveniently as it was not intended.

I have explained why it is necessary to have a fixed point! As things are and as the event was explained a train was sent towards the single track section trapping the train that was already in the section. Think of the track as a wishbone, one road in and the other road out with the points at the junction of the two, when a train is too close to the points on the in road it fouls the out road because there is not room for a train to pass or as in this particular case locked the points. Thus an exceptional move (setting back) had to be made to free the points.
Proper signalling would have halted the approaching train before it trapped the one in the platform.

[Deleted]

Sept 16, 2013 21:04:07 GMT railtechnician said:

I have explained why it is necessary to have a fixed point! As things are and as the event was explained a train was sent towards the single track section trapping the train that was already in the section. Think of the track as a wishbone, one road in and the other road out with the points at the junction of the two, when a train is too close to the points on the in road it fouls the out road because there is not room for a train to pass or as in this particular case locked the points. Thus an exceptional move (setting back) had to be made to free the points.
Proper signalling would have halted the approaching train before it trapped the one in the platform.

No different to a signaller inadvertently clearing WB3 at West Kensington, allowing a train into the Westbound platform with a train waiting to shunt east at WC31 and an engineering shutdown west of WC7 (as happened in the not too distant past). The outcome is the same - a train stuck on a (de facto) single line because another is (quite legitimately; safely, and with a safe separation distance maintained between it and the other train) sitting on the points which it needs to proceed over to vacate the single line. A wrong direction move then has to take place.

As long as the TBTC system knows that it must not route a train towards Mill Hill East if one is already on the branch, and that the operator may only initiate such a move after confirming their intention then it is simply not an issue IMHO and I'm yet to be dissuaded from that view.