DAP 2009 Q4: Stopping Trains in Emergency
One of a set of three questions undertaken in mock exam conditions
============================================================================================
Part a) option 1
Careful with wording “only affects one signal”- obviously every signal within aspect sequence range reverts accordingly; it may have been worth making more of this by stating if train far enough away then it should stop at the protecting signal normally but is otherwise likely to SPAD, dependent upon its position when the driver first becomes aware of the more restrictive sequence, its speed and rail head conditions etc. In the case of continuous ATP then the information can be conveyed to the driver before the changed signal itself becomes visible; indeed in this scenario the driver is actually likely to be alerted to the situation by the braking initiated by Automatic Train Protection.
Part a) option 2
ESOC is not necessarily implemented via the crude method of cutting the power to the central interlocking; for example a WESTLOCK cubicle has a directly wired input and if this feed is lost the OPT file is no longer processed (so all signals go to red) but the indications remain valid and normal working can be resumed immediately once input is restored. The essence of the control is
i) a single quick action to revert all signals within a defined area (as you stated)
ii) a means that does not rely upon the operation of the VDU control system in order that there is a diverse means to place the signalling into a safe state even should this freeze up or become inoperative for some other reason (indeed it was when these systems were a novel replacement for the trusted panel directly wired to a relay interlocking that it was felt that there was a need for ESOC).
Your text was fine for SSI and SMARTLOCK and indeed it gave you more to discuss- it would however have been better if your answer had recognised the distinction between the “concept” and “particular implementation”.
Part a) option 3
I admit that I am not sure of the details of the GSM-R Global Emergency Stop; I suspect you are correct that there is no positive feedback of receipt as it is a broadcast call to anything that happens to be in the selected area (normally a particular GSM-R cell and all of its neighbours, although this is configurable) and indeed can be initiated by any driver as well as the signaller. I don’t believe that there is any direct tie up, but for ETCS L2 (and above) the train position and speed is reported back to the RBC; on the Cambrian lines the signaller has a supplementary display provided primarily for the imposition/ removal of Temporary Speed Restrictions on which the train is depicted at its longitudinal position compared to the signalling plan and supplementary details of its status, such as speed, can be viewed.
Despite comments above, part a) was basically well answered and of appropriate length. Perhaps you could have said a little more about effectiveness and particularly for option 3 have included more discussion regarding the area affected. Being based on radio cell coverage it is inevitable that there will be instances of far more trains being affected than ideally would be. The primary problem is when there are two operationally all but independent railways in the same radio cell and therefore two logically separate routes are interconnected; for example places such as Tamworth where one mainline crosses another at a different level, places such as Willesden/Old Oak Common where completely separate lines actually run extremely close together and radio propagation effects mean that some places on a line might have better coverage from a cell for the other line, place such as Richmond where the terminal platforms of the London Overground/ District Line are operationally nothing to do with the through platforms on the line to Waterloo.
Even without such issues, the area of a railway within a GSM-R cell is unlikely to match precisely those signals which would be chosen to be replaced in an emergency. Typically a Railway Emergency Call is given to all trains within not just the cell but also to all neighbouring cells in an attempt to stop trains from getting into the danger area- however this can mean that the effect is quite widespread. Conversely a definite safety advantage is that all drivers become immediately aware even if between signals and also that the REC can be initiated by any driver from their cab; this can of course have some disadvantages; there was an instance not long ago of a telecoms technician attending to a fault on radio equipment in a tamper repeatedly bringing the operating railway to a stand whilst installing a replacement piece of equipment in its cab and being blissfully unaware of the chaos being created by their antics.
Part b)
Good answer, but it would have been worth bringing into the discussion the likelihood of signallers being faced with different arrangements within the disparate areas all gathered together within a large Rail Operating Centre- the signalling technology hat has been re-controlled to it and indeed the geography of the various lines within a large operating area could be very different. There are still a lot of individual signalboxes left extant and certain signallers may be passed out for many which could be quite different from each other and which they actually operate very rarely (e.g. relief signallers, MOMs). The RAIB have recently issued an accident report into the derailment at Knaresborough which relates to a signallers unfamiliarity which was the principal cause of the derailment, but also the driver not selecting correctly the REC in the immediate aftermath and therefore not achieving the level of mitigation that should have been possible; both elements are salient here.
Part c)
As I understand it, when using GSM-R, all emergency stop communication must be to all trains within a specific radio cell and those neighbouring cells which have been pre-defined in the design as associated with it. This is both to get a good speed of response and to maximise the probability of driver reaching the message.
Whereas the radio system certainly allows a normal call to be routed to a particular train, this does rely on accessing several central databases (to associate an entered train description with the train at a particular signal, to associate that train identity with the particular rolling stock and indeed the currently being used cab within that formation, to associate this cab with a particular SIM card in order to know what number to ring). Clearly obtaining this information takes time and if any of the requested information is absent or invalid, then the call would not succeed; in addition there is unlikely to be enough capacity within the cell to simultaneously initiate separate calls to all the different trains which might be within it. Hence the approach taken is the broadcast emergency call to anything within the area; it is not reliant upon the driver correctly having completed registration etc.
Whereas I thought your answer was broadly reasonable generically, I don’t think it fully reflected the specific of GSM-R functionality in the UK (one could argue that it didn’t need to and I am certainly also out of my comfort zone here.)
Overall though I felt this was n answer that would be a Credit.
One of a set of three questions undertaken in mock exam conditions
============================================================================================
Part a) option 1
Careful with wording “only affects one signal”- obviously every signal within aspect sequence range reverts accordingly; it may have been worth making more of this by stating if train far enough away then it should stop at the protecting signal normally but is otherwise likely to SPAD, dependent upon its position when the driver first becomes aware of the more restrictive sequence, its speed and rail head conditions etc. In the case of continuous ATP then the information can be conveyed to the driver before the changed signal itself becomes visible; indeed in this scenario the driver is actually likely to be alerted to the situation by the braking initiated by Automatic Train Protection.
Part a) option 2
ESOC is not necessarily implemented via the crude method of cutting the power to the central interlocking; for example a WESTLOCK cubicle has a directly wired input and if this feed is lost the OPT file is no longer processed (so all signals go to red) but the indications remain valid and normal working can be resumed immediately once input is restored. The essence of the control is
i) a single quick action to revert all signals within a defined area (as you stated)
ii) a means that does not rely upon the operation of the VDU control system in order that there is a diverse means to place the signalling into a safe state even should this freeze up or become inoperative for some other reason (indeed it was when these systems were a novel replacement for the trusted panel directly wired to a relay interlocking that it was felt that there was a need for ESOC).
Your text was fine for SSI and SMARTLOCK and indeed it gave you more to discuss- it would however have been better if your answer had recognised the distinction between the “concept” and “particular implementation”.
Part a) option 3
I admit that I am not sure of the details of the GSM-R Global Emergency Stop; I suspect you are correct that there is no positive feedback of receipt as it is a broadcast call to anything that happens to be in the selected area (normally a particular GSM-R cell and all of its neighbours, although this is configurable) and indeed can be initiated by any driver as well as the signaller. I don’t believe that there is any direct tie up, but for ETCS L2 (and above) the train position and speed is reported back to the RBC; on the Cambrian lines the signaller has a supplementary display provided primarily for the imposition/ removal of Temporary Speed Restrictions on which the train is depicted at its longitudinal position compared to the signalling plan and supplementary details of its status, such as speed, can be viewed.
Despite comments above, part a) was basically well answered and of appropriate length. Perhaps you could have said a little more about effectiveness and particularly for option 3 have included more discussion regarding the area affected. Being based on radio cell coverage it is inevitable that there will be instances of far more trains being affected than ideally would be. The primary problem is when there are two operationally all but independent railways in the same radio cell and therefore two logically separate routes are interconnected; for example places such as Tamworth where one mainline crosses another at a different level, places such as Willesden/Old Oak Common where completely separate lines actually run extremely close together and radio propagation effects mean that some places on a line might have better coverage from a cell for the other line, place such as Richmond where the terminal platforms of the London Overground/ District Line are operationally nothing to do with the through platforms on the line to Waterloo.
Even without such issues, the area of a railway within a GSM-R cell is unlikely to match precisely those signals which would be chosen to be replaced in an emergency. Typically a Railway Emergency Call is given to all trains within not just the cell but also to all neighbouring cells in an attempt to stop trains from getting into the danger area- however this can mean that the effect is quite widespread. Conversely a definite safety advantage is that all drivers become immediately aware even if between signals and also that the REC can be initiated by any driver from their cab; this can of course have some disadvantages; there was an instance not long ago of a telecoms technician attending to a fault on radio equipment in a tamper repeatedly bringing the operating railway to a stand whilst installing a replacement piece of equipment in its cab and being blissfully unaware of the chaos being created by their antics.
Part b)
Good answer, but it would have been worth bringing into the discussion the likelihood of signallers being faced with different arrangements within the disparate areas all gathered together within a large Rail Operating Centre- the signalling technology hat has been re-controlled to it and indeed the geography of the various lines within a large operating area could be very different. There are still a lot of individual signalboxes left extant and certain signallers may be passed out for many which could be quite different from each other and which they actually operate very rarely (e.g. relief signallers, MOMs). The RAIB have recently issued an accident report into the derailment at Knaresborough which relates to a signallers unfamiliarity which was the principal cause of the derailment, but also the driver not selecting correctly the REC in the immediate aftermath and therefore not achieving the level of mitigation that should have been possible; both elements are salient here.
Part c)
As I understand it, when using GSM-R, all emergency stop communication must be to all trains within a specific radio cell and those neighbouring cells which have been pre-defined in the design as associated with it. This is both to get a good speed of response and to maximise the probability of driver reaching the message.
Whereas the radio system certainly allows a normal call to be routed to a particular train, this does rely on accessing several central databases (to associate an entered train description with the train at a particular signal, to associate that train identity with the particular rolling stock and indeed the currently being used cab within that formation, to associate this cab with a particular SIM card in order to know what number to ring). Clearly obtaining this information takes time and if any of the requested information is absent or invalid, then the call would not succeed; in addition there is unlikely to be enough capacity within the cell to simultaneously initiate separate calls to all the different trains which might be within it. Hence the approach taken is the broadcast emergency call to anything within the area; it is not reliant upon the driver correctly having completed registration etc.
Whereas I thought your answer was broadly reasonable generically, I don’t think it fully reflected the specific of GSM-R functionality in the UK (one could argue that it didn’t need to and I am certainly also out of my comfort zone here.)
Overall though I felt this was n answer that would be a Credit.
PJW