(16-06-2010, 01:08 PM)Hort Wrote: I have had another go at the question.
Any comments would be appreciated.
You are certainly considering the correct scenario this time.
You have some appropriate hazards and have values for likelihood and severity, then calculated the risk scores.
You have a list of mitigations and related these to the hazards and calculated revised risk rating, so from that point of view I thought it was great.
I wasn't convinced that it made sense to calculate "risk change"; it suggests that the value has actual numerical significance rather than just being a relative weighting.
I think you must have been assuming (but never stated) that each mitigation was considered in isolation. Take hazard 3 for example: original likelihood 3. You have three mitigations: one of which actually makes no significant change, the second reduces severity and the third reduces likelihood. Obviously if you were implementing them all then you'd have a reduction both of likelihood and severity and only have one revised score for the hazard.
My main concern with your answer was that the question scoring is
Carry out a risk assessment of all aspects of the operation including maintenance and the operation of infrequent wrong direction movements. [16 marks]
Briefly identify possible mitigations for the risks that you identify. [5 marks]
You gave more on the mitigations (it said "briefly identify", 5 marks) than the risk assessment (it said "carry out", 16 marks).
I think that your table should have been simplified by making the second column more of a list (certainly no more than 2 lines of text) and just an indication in the next two columns whether it was to rduce likelihood or severity or both. I think giving revised scores wasn't called for, so save the time. You couldn't have done this page in 6 minutes I think!
Conversely you ought to have spent about 16 minutes on the risk assessment, so the examiners were looking for more. I think that I would have done it much like you, but where there was a hazard such as 1 = SPAD then I'd have split that horizontally into 1a and 1b, the first when reversible working is not in use and the second when it is. This would make it clear that you were considering the two situations.
You say for item 5 that the risk to maintenance staff is higher when using a line in the wrong direction is increased; I agree with you that it is whilst that situation exists. Let's say that it is increased tenfold; however is it likely that reversible routes will be in use for as much as 10% of the time? So which is the biggest risk over the period of a year?
I think that you had a reasonable range of hazards - lack of adhesion, loss of train detection, driver overspeeding, driver failing to respect stop signal, maintenance workers inadequately protected from train movements etc. so what other ones could you have included?
The bit of the question that you haven't incorporated is No train protection system is provided for the reverse direction movement nor is there any suppression of the right direction
train protection.
I do have a bit of a problem with this, thinking in NR context. Obvious form of train protection is TPWS- the frequencies of the loops makes this inherently directional so a wrong direction train won't get tripped- it doesn't need "suppression". Similarly Eurobalise, there is an inherent directionality between the two in a balise group. I am not an expert on the various ATPs but I believe the beacons are offset to the track centre line and they get directionality in that way and again don't need suppession. That only leaves AWS which I don't regard as train protection; I think I need to state this as an assumption to the examiners but then consider the lack of AWS suppression. So the hazard is that the driver gets used to cancelling the bogus AWS warnings whilst travelling on the wrong line and then carries on doing so when regains the right line.
I think that it is reasonable to assume that a line like this may have some level crossings, so perhaps that gives another hazard we can include.
I'd have made more of the loss of train detection. If there is a track bob then we could loose all the rout locking ahead of the train and a route could be set in the opposite direction. The most likely scenario is whilst train traversing those crossovers that may be very rusty from infrequent use. Mitigations might be eucleptic welded strip on the crossover rail surface and specially slow-to-release route locking for the overall totality of each reversible section (proving all track circuits for whole length between crossovers simultaneously clear for say 10 second, before permitting reversal of directionality of use).
We could perhaps consider where the signals for the wrong direction signal would be situated; perhaps they are straight posts on the righthand side of the line rather where the driver would normally expect to see them, so perhaps this would increase the risk of them being SPADed.
Could it be that because of line curvature that a driver on a train running on the normal line might see the wrong road signal first? Given that the ones approaching crossovers are only Y/G signals then there default aspect is yellow which is a proceed aspect. Driver who had receiverd Y on previous signal might think that the signal he is approaching has since cleared and cceases braking but then the R comes into view and travelling too fast to stop at it.
What about the risk of a driver travelling in the wrong direction sees a Green on the protecting signal for the crossover but failing to appreciate that it is for regaining the correct running line and therefore encountering the points too fast. Yes you had considered overspeeding here as a risk, but not really this scenario.
Other than that I can't think of other obvious omissions specifically to do with reversible working; the question certainly asks for operation in normal scenario as well but since is bascally a two line uni-directional railway then there really isn't much- a few causes of a SPAD leading to a rear end collision and you have these.
Perhaps we could include things like track failures, points failure so that we are also considering degraded mode working. Perhaps we could extend this to consider planned engineering works on ne line with a reduced train service on the other. Basically I am struggling now to get more for this section; I don't think I have missed anything so I guess that the examiners are wanting this sort of thing.
Overall this was a reasonable attempt,
however-
far too much intro for which there would be few marks,
very light on the miiddle section where the majority of the marks are, and
more than you could spare the time for in the last section
------------------------------------------------------------------------
A small point (at least re this question and this module is concerned). Some of what you wrote re signalling principles etc was a bit dodgy. I don't think it would have harmed you significantly in gaining marks in this question, but in other cases it would. [e.g. Double Red- the outer signal (you said "in advance" when you meant "in rear" or as we should say now "on the approach" clears once the train has passed beyond its OSS and therefore the train may be some 400m prior so certainly not at a stand]
PJW