2012 Q7 Signal Sighting

[PJW]

Attempt in exam conditions, took a little over 30mins.

I'd not read the previous attempt nor the input from PJW, so I can see where I should have added more detail.

Any feedback is welcomed. I'm not sure where the 6 marks come from in part c.

Regards,
Adrian

a) Did plenty for the first 4 marks, but it would have been good to have suggested some of the underlying reasons which may have led to problems in the visibility that you have described (you didn't really state much in the way of assumptions)-

• that the footbridge may be obstructing just the route indication above the signal, particularly if the train cab is some distance on the approach to the footbridge which cuts off field of view,
  
• increased crash resistance of modern rollingstock tends to mean that the driver is positioned further back from the front window and this restricts peripheral visibility (particularly an issue for ground mounted signals- although this not relevant in this case)
  
• trains may not always be run in full length formations and it is when a short train is at the buffer stop end that platform curvature may be most of an issue,
  
• right hand mounted signals can be particularly hard for a driver to see from their seat on the left of the train when cab is close to signal and this may be constrained because of the desire that the new trains are as long as possible for maximum capacity.
  

b) You hadn't stated a context for your answer- perhaps it didn't matter for part a, but I think you should have stated before answering part b as the parties involved for London Underground differ somewhat than those on NR.  Would there not be any representative for the traction- they may not have any responsibilities re the signal visibility per se, but presumably there are likely to be conductor rail implications for fitting the trainstop etc.


c) Be careful of wording re what hazards exist;
a derailment or collision counts surely as an ACCIDENT since there is a certainty of loss- there may not be death and perhaps even injury might not result, but there would certainly be property damage and service disruption at the very least and that is a definite loss, not a potential one!

Conversely a signal overrun is a hazard; there may or may not be a train routed over the crossover at that time  and so there is a potential for collision and the level of risk associated with a head-on collision is rather more than a rear-end or side-swipe,
similarly there is the potential to run-through the trailing points depending on which lie they happen to be, overspeed is a hazard since derailment is a possible (although in this case probably quite unlikely) outcome,
any overrun will bring an element of service disruption, but the extent of this does depend on the level of traffic at the time and the positions of other trains, so whether significant harm is done does indeed depend on a range of factors and thus is quite uncertain.

d) I agree- think getting these 6 marks would be very difficult- your possibly solutions seem the obvious ones; however don't see that what you wrote was worth more than 4 at most, nor really what else you could have added.  I suppose that you could have pointed out that the repeater of no value if the sighting issue was one resulting from front cab close proximity to signal (I may be thinking too much in a mainline context for your answer- a banner repeater would almost certainly go on or on the approach to the footbridge, but if needed the option of a co-acting signal could be provided either ground mounted and/or on the left of the line- I guess that an LU repeater could actually be used in either of these scenarios)

e) This was indeed expecting you to describe the 7 stage process.
Yes could use a Fault Tree to consider the likelihood of the Top Event of SPAD- this would get a combined estimate from the various scenarios which could lead to that occurrence.
Not convinced however that FMEA would be appropriate in such a case; we are really considering the chance of human error here, so we could get some values using the HEART technique re for e.g 

• the likelihood of the driver failing to check the signal before departure, perhaps triggered by hearing the points move, closing the doors and then having checked the interlock and checked the CCTV just powering away.
  

Then would use an Event Tree to model the possible consequences, based on an assessment of the train protection system (remember that at this signal there is no locked overlap within which a SPAD can safely be contained), the possibility of the driver becoming aware by other means, the timetable based estimate of how likely it is that there is another train in the area to collide with in a certain manner (head-on, sideswipe, rear-end) and therefore the total estimate of loss, based upon the proportionally summated probable consequences of each of the potential outcomes.

These would be done for the original unmitigated state, then re-done making alterations to reflect your two proposed options and therefore values obtained for the original level of risk and the two others.  You should have tied your "options analysis" back to the earlier parts of this question.

The important final thing to make clear is-
1. Risk must not be intolerably high
2. Risk should be reduced to be ALARP and to be able to demonstrate that then the cost of each of the mitigations must be compared to the safety risk reduction it is estimated to have.
3. Whereas a simple cost-benefit analysis is appropriate to non safety risk (is it worth the investment to avoid a possible financial loss?), the assessment of a mitigation of a safety risk is subtly but significantly different (having checked that that the residual risk here is itself tolerable, do any of the options represent the best use of the safety £ or would money actually be more effectively spent elsewhere to get a bigger safety improvement there?)