Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Is approach release fail safe?
#1
Hi guys,

An interesting debate on "approach release"......

We approach release the signals to reduce the train speeds to a reasonable extent. But.....

"Do you think approach releasing the signals is fail safe?"

Regards,
Saraswathi
Reply
#2
(12-01-2010, 07:57 PM)Saraswathi Wrote: Hi guys,

An interesting debate on "approach release"......

We approach release the signals to reduce the train speeds to a reasonable extent. But.....

"Do you think approach releasing the signals is fail safe?"

Regards,
Saraswathi

Yes this could be the basis of a mod3 written question. In order to give enough scope for a 30 min answer then I think that I might word it:
Describe 3 situations in which the clearance of a signal is deliberately delayed until there is a train on its approach.
For each situation-
a) explain the rationale for imposing the control,
b) comment on any safety implications,
c) describe how any risks are mitigated.

Let's see some attempted answers to this.
PJW
Reply
#3
(12-01-2010, 08:42 PM)PJW Wrote:
(12-01-2010, 07:57 PM)Saraswathi Wrote: Hi guys,

An interesting debate on "approach release"......

We approach release the signals to reduce the train speeds to a reasonable extent. But.....

"Do you think approach releasing the signals is fail safe?"

Regards,
Saraswathi

Yes this could be the basis of a mod3 written question. In order to give enough scope for a 30 min answer then I think that I might word it:
Describe 3 situations in which the clearance of a signal is deliberately delayed until there is a train on its approach.
For each situation-
a) explain the rationale for imposing the control,
b) comment on any safety implications,
c) describe how any risks are mitigated.

Let's see some attempted answers to this.

Hi Saraswati

Please find attached document which may probably answer your question to some extent. I think it requires some more inputs (mainly from UK Main line front). Can somebody analyse this further?

Nagaraju


Attached Files
.doc   APPROACH CONTROL.doc (Size: 56.5 KB / Downloads: 158)
Reply
#4
The reason why i have started this debate is:

The concept of "Approach control" is to mainly reduce the train's speed to a certain extent by keeping the signal at red/yellow. But this is achieved and mainly based on its berth track occupied condition...
What if the berth track is failed in the de-energised condition? That can easily turn the approach controlled signal to get cleared which is not fail safe.

Any more views on this? Or plz correct me if i am thinking wrong....

Regards,
Saraswathi.
Reply
#5
I may be wrong but surely "approach control" is an aspect condition which is based on a valid route being set. The signal will not clear from a red aspect due to the <<berth>> track not proven clear. Remeber, App Cont is usually berth OCC plus 't secs' [time].

Therefore, you are still fail-safe under "normal" failure conditions. There are credible but improbable conditions where the berth track was picked under specific circumstances that could allow the aspect to clear to a better aspect earlier than was designed. However, you have a further mitigation that the driver has route knowledge and should be preparing for the speed restriction [in the UK].

Jerry
Le coureur
Reply
#6
If the track circuit that controls the release of the aspect is failed occupied, remember that it is one of the tracks in the line of route of the signal in rear, so the signal in rear will not clear up to the approach released signal if this failure condition is present.

Of course you could have the TC operating normally and when the train has gone past the signal in rear, the TC MAY fail as the train comes along and release the signal early, but you will have had the yellow in the signal in rear.
Reply
#7
In London Underground, they use Delta Rail Circuit for approach control. Delta Rail Circuit is a 10kHz open circuit superimposed on existing AC track circuit,whose length is small. The Delta Circuit is closed when shunted by wheels of train. The exact location of train on the approach to a signal can be verified by energisation of delta relay and the front contact of this delta relay is used to clear the signal. In this way this method used on LU is fail safe because failure of delta relay means signal will not clear. But coming to UK Main line signalling there appears to be different techniques used for approach control, each having its own merits and de-merits which I need to investigate further.


(05-03-2010, 04:48 PM)Peter Wrote: If the track circuit that controls the release of the aspect is failed occupied, remember that it is one of the tracks in the line of route of the signal in rear, so the signal in rear will not clear up to the approach released signal if this failure condition is present.

Of course you could have the TC operating normally and when the train has gone past the signal in rear, the TC MAY fail as the train comes along and release the signal early, but you will have had the yellow in the signal in rear.
Reply
#8
nagaraju_dw,

To the best of my knowledge for approach control, we use delta tracks (or position detectors or track circuits) to start a timing relay (JR) which proves the speed of the approaching train is appropriate when only a restricted overlap is available at the following signal. If the test is passed a timing stick relay (JSR) is energised. The JSR is then included in the signal relay (GR) circuit. The delta relay and JR are down proven in the previous signal's selection.

Alasdair
Reply
#9
(05-03-2010, 03:24 AM)Saraswathi Wrote: The reason why i have started this debate is:

The concept of "Approach control" is to mainly reduce the train's speed to a certain extent by keeping the signal at red/yellow. But this is achieved and mainly based on its berth track occupied condition...
What if the berth track is failed in the de-energised condition? That can easily turn the approach controlled signal to get cleared which is not fail safe.

Any more views on this? Or plz correct me if i am thinking wrong....

Regards,
Saraswathi.

Raj has given an interesting answer that will be worth discussing separately, but he hasn't actually addressed the issue that was worrying you.
You are absolutely correct that a track circuit is designed to detect (in as much of a fail-safe manner as possible) the ABSENCE of a train. Hence it does not make a very satisfactory means of reliably detecting the PRESENCE of a train; this is ideally what we'd have to implement approach release.

A failure will indeed result in the false release of the aspect- a wrongside failure. Risk is that driver gets a clear aspect sequence on the signals approaching the junction and by the time they see the route for which the points are set, may well be unable to slow the train to a safe speed to traverse the divergence safely and may overturn / derail.

In essence what we do however is make this a PROTECTED WRONGSIDE failure. We can't stop the false release of the junction signal occuring (unless we utilise a differnt means of train detedction perhaps); we can very significantly mitigate the risk by knowing that it has happened and holding the previous signal to red.

The name that is now frequently given is "Raynes Park control"- named after the place n south London where the signalling circuits were in fact designed incorrectly and the accident occurred.
The rule is that whatever is used to perform the approach release (may be a track timer, a track circuit, a treadle being depressed) is proved in the opposite state within the controls of the signal in rear.

Suppose track AE is required occupied for 25 seconds to undetake the MAR on the junction signal S5. If the track circuit fails rightside; signal S5 will then soon show perhaps Green and PLJI1 route indicator. However by down-proving this approach release within the controls of signal in rear, S3, then this signal that had previously been displaying yellow will not step up to Green (assuming 3 aspect) but instead revert to Red. Train will thereforee be brought to a stand at it and driver will speak to signaller and then be authorised through the section at caution and contact again at the next signal. Hence although driver does see S5 at the invalidly released aspect, there is no risk as the train is only travelling slowly and has been told to stop to report.

There is a residual risk that the track fails at just the wrong time; the train has already passed S3 when it fails. In this case S3 will have been seen at Yellow so train is braking expecting to stop at S5 but driver could then perhaps see S5 in the far distance at Green and accelerate for a whilde before becoming aware that it is really at Green and PLJI1 and therefore for the speed restricted route. In this scenario the mitigation is not 100%; however remember:
a) the likelihood of the failure happening is low- there is perhaps a 60 second "time window" for the average signal during which the track could fail and create this undesireable result- any earlier signal S3 would revert before train passes it, any later and the train will be in the area where it would be dropping the track anyway
b) the consequences are reasonably low. Unless it is a very low speed divergence from a very fast running line, the degree odf overspeeding will be within acceptable limits- the ride may be uncomfortable but the train is most likely to stay on the track, so perhaps the incident will be someone walking down the train with some coffee spilling scalding liquid over a few other passengers- not multiple deaths from train thrown off track and smashing into OHLE masts or plunging into river.

Clearly the risk is worst if the main aspect of the signal is visible from a very great distance. the train has good acceleration and the PLJI only becomes visible very close to the signal. Perhaps signal spacing 1600m, good straight track so Green from S5 can be seen at 1000m but due to an overbridge blocking the PLJI from view then this remains invisible until 300m. Acceleration rate likely to be around half the braking rate so even in this extreme scenario train likely to be doing less than max permissible speed when site the PLJI and there is still some distance available to make a further speed reduction- indeed the points themselves are quite often at least an overlap length beyond the junction signal, so could easily be 500m from visibility of PLJI. Only need to reduce speed of train to be less than the actual derailment speed- pretty likely to "get away with it" at some 30-50% higher speed than the nominal design speed for the divergence.

Of course the track itself is naturaly in the controls of the signal in rear- we just have to be careful that it really is correct generation repeat relay. The think to be careful of is if the TPR that is used in signal S3 is up yet the T2PR repeat relay used for approach release of signal S5 fails due to a blown fuse then there would be a problem. Solution is either tio use a back contact repeat of the TPR to perform the approach release of S5 or ensure that it is the T2PR that is used in signal S3. Similarly if a track timer is used for approach release, then this must be proved to be in its quiescent (untimed) state for the signal in rear to clear; conversely if a treadle being depressed is used for approach release then it must be proved normal (not pushed) within the controls of S3.

The above is not a full answer to the question I set (and indeed there are other elements to bring in from Raj's document) but hopefully it goes some way to addressing the issue and responds to your original question which I thought pertinent and demonstrated that you are really thinking about the subject to have raised it as a topic for debate.
PJW
Reply
#10
(07-03-2010, 05:42 PM)Alasdair Wrote: nagaraju_dw,

To the best of my knowledge for approach control, we use delta tracks (or position detectors or track circuits) to start a timing relay (JR) which proves the speed of the approaching train is appropriate when only a restricted overlap is available at the following signal. If the test is passed a timing stick relay (JSR) is energised. The JSR is then included in the signal relay (GR) circuit. The delta relay and JR are down proven in the previous signal's selection.

Alasdair

You are correct. Delta relay pick will be proved in timer circuit (JR/JSR) and JSR pick will be proved in signal relay circuit.
Thanks for the correction Alasdair.

Nagaraju
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)