Overlap sharing

[Saraswathi]

Can anybody explain:

What is overlap sharing?
When can we share the overlaps and when we can't?

[PJW]

Can anybody explain:

What is overlap sharing?
When can we share the overlaps and when we can't?

An overlap is the safety overrun distance that is reserved beyond the route's exit signal, in case a train fails to stop.
Generally in the UK then it is for it's exclusive use, just in case.

However SPADs are actually quite rare and therefore the risk of two simultaneous SPADs in the same area is pretty minimal (disregarding what we call cat B or "technical" SPADs when the signal concerned was legitimately cleared but reverts due to some fault such as loss of point detection or a TC failure. Although driver of train will pass the signal at danger, it is not their fault, nor is there risk since the route was stil set and locked for that mvement due to approach locking / route locking. So two cat B SPADs in an area could well be caused by the same fundamental fault, but zero risk of train collision).

So we could risk assess and say that two signals could share overlap.

This may be "head on"; on a bi-directional line an Up signal might have a 180m beyond it that terminates at the block joint at the Down signal position. If overlaps are not shared clearly there could not be a route set up to that Down signal as it could not have overlap, since reserved for the Up movement. However we could dcide that they could share on the basis that a SPAD at both at same time is not credible.

It could also be "converging". Imagine a double track section of passing loop in a station area becoming a single track with trailing points to the single line portion just beyond loop exit signal.
If an Up train is arriving at loop from double track section, then it would certainly not be sensible to permit a Down through train to pass via the other loop simultaneously, since if first train failed to stop would almost certainly be immediate head on collision. However we could decide that we'd be prepared to allow two Up direction movements, one into one loop from the rnnung line and the other into the other loop from sidings. If either train did SPAD (credible risk) then it wouldn't hit the other.

Generally in the UK we do not share overlaps, but certainly the Western Region did where line speed was low and risks felt reasonable (indeed Reading's new WESTLOCK interlocking commissioned in Dec 2010 implements those principles whilst the existing track layout remains, so it is still something that we do in particular circumstances). Indeed with "risk assessment" to the fore and the traditional sometimes rather black/white rules being revisited, it is certainly possible that such policies will be implemented more widely.

So the question you asked would be just the sort of thing that might come up as a written question in 2011 Exam; anyone feel like writing an answer?

Sara, from your reading of various Australian practices, do any railways there share overlaps?

Any comments from others from "down under" or other parts of the world?

[Saraswathi]

Thanks for the detailed response Peter. This concept came into light when I read about this overlap sharing in the IRSE news magazine of January (article on NSW signalling principles) where it was mentioned that New South Wales (which is one of the Australian state here) do not use shared overlaps. I have then looked at other state's standards and interestingly found that Queensland state permit overlap sharing. But they have a condition which could probably mitigate the risk to certain extent. QR (Queensland Rail) standard states that: "Overlaps may be simultaneously shared provided that at least 50% of each overlap is available before any point of conflict (e.g. head on collision or clearance infringement). In addition there must be a minimum of 50 metres available for each approaching train before any point of conflict."

Regards,
Saraswathi Penneru

[Saraswathi]

Peter,

Shared overlaps, are they equivalent to phantom overlaps? I was going through one of discussion in this forum and found the topic on Phantom Overlaps and the reply given by you, where overlap sharing is also discussed:

http://www.irseexam.co.uk/showthread.php...om+overlap

It says "We are not allowed to "share overlaps" but we can "share overlap track circuit" by defining a POL in the middle. The aspects of the two opposing direction routes up to the signals prove the entire 360m clear; however even if both trains then SPAD by 180m then they won't actually hit each other and it is just as safe (arguably more so as automatic reversion of a valid aspect if the other train SPADs) as if separate TCs provided."

Regards,
Saraswathi Penneru

[PJW]

Peter,

Shared overlaps, are they equivalent to phantom overlaps?

Regards,
Saraswathi Penneru

Not quite. I was actually intending to add to my previous post a diagram to make things clearer and was already thinking of including phantom; I just need to grab some time to do this.

For now though, just words:
If there are 2 opposing signals (113 and 114) and we wish to signal up to each simultaneously, we would need each to have its own overlap distance, say of 180m reserved for its own unique use (assumption of NOT permitting overlap sharing). However we need not necessarily provide a track circuit boundary there; we could make the overlaps both indicate as one track. So instead of tracks AB and AC we would just have track BB of twice the length it would have been otherwise.
This means that occupancy anywhere between the signals would replace both and there is a sharing therefore of the train detection (and thus saves some cost) but if there were two SPADs each could overrun to the end of its own overlap and not actually collide, other than the most gentle touch of the buffers perhaps.
The safety is precisely the same as if two 180m track circuits had been provided, but economy achieved by providing one 360m track instead. Actually arguably slightly improved since, as soon as the first SPAD occurs at 113 and BB becomes occupied, the outer opposing direction signal 116 (i.e that authorising up to 114 which immediately protects) would revert to danger.
In the separate overlap TC instance, track AB would not be naturally proved within the opposing direction overlap so that signal would only revert if the SPAD exceed that length and got as far as AC. Of course nowadays we may well also specifically provide "overrun detection" ; a special control that replaces relevant aspects in the area if a SPAD is detected (e.g. if AB becomes occupied after AA occupied whilst the protecting signal 113 at danger) then the interlocking or control system would take emergency action to mitigate the risk by attempting to stop valid movements in the vicinity.

The difference with a shared overlap is that the signals need only be 180m rather than 360m apart- not only is the train detection section shhared but so also is the overrun distance. It is less safe since the likelihood of collision has been increased ( very very marginally- do you actually consider that it is credible that the two SPADs would occur more or less simultaneously? Once one has occurred it will revert the outer signal so we are looking to risk assess how likely it is that there would at that moment be a train coming which itself wasn't actually going to stop where it was originally supposed to).

When I have time to add diagram, hopefully it should make it clearer.

[PJW]

Peter,

Shared overlaps, are they equivalent to phantom overlaps?

Regards,
Saraswathi Penneru

When I have time to add diagram, hopefully it should make it clearer.

Now done; read with above post