GC class B1
Member
Could you please elaborate on what is meant by speed allowed according to the braking curve.
-
Our new ticketing site is now live! Using either this or the original site (both powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!
Derailment at Kirkby (Merseyside) - 13/03/2021
- Thread starter tgsh2011
- Start date
- Status
Sponsor Post - registered members do not see these adverts; click here to register, or click here to log in
R
RailUK Forums
TPWS "mitigates the circumstances of" but doesn't "prevent"
That shouldn't have happened. I'm sure if you let them know on enquiries@raib.gov.uk a suitable response will be forthcoming.Incidentally, it's nice that my photo of an AWS cancelling indicator is used in the report (page 18), although I wasn't asked for permission and no acknowledgement is given.
Stuart J
RAIB (retired)
zwk500
Veteran Member
A braking curve is the progression of speed allowed when approaching a Red Signal or buffer stops in order to stop safely. So it'll start at linespeed and trend towards 0 as you get closer to the red signal. As its the result of various equations, on a graph the line is curved. There are all sorts of modifiers for gradient, local circumstances etc. The TPWS activation speed is set to the appropriate value according to the braking curve for it's position relative to the signal.
DanNCL
Established Member
If possible I would change the one on the approach to the PSR so that such a high speed upon reaching the one in the platform isn’t possible. If that wasn’t practical then I’d install a third one between the two existing ones.
It doesn’t even seem to have mitigated in this case.
GC class B1
Member
I agree completely.
Imagine if we didn't have TPWS and the Driver ploughed into the buffers at full speed, whilst stil accelerating...
GC class B1
Member
Thank you. Can you please let me know what deceleration rate is used to determine the 53MPH set speed for the 15 MPH permanent speed restriction OSS loop.
zwk500
Veteran Member
I think you've got confused. There are two separate TPWS units. The first trips the brakes at 53mph. The second, later on up the track, trips the brakes at 15mph. The driver was travelling at lower than 53 mph when he passed the first TPWS grid, and should have been able to stop safely. However, for reasons outlined in the report, he did not continue to brake until he realised he was too fast and pulled the emergency brake. He had already pulled the brakes before he reached the second tpws unit.
I also can't provide the equation because I no longer have access to the relevant NR or RSSB standards, but somebody else may be able to, if you ask them nicely.
DanNCL
Established Member
TPWS did nothing in this case, as the driver had already applied the emergency brake when the train reached the second TPWS loop.
Thank you. Can you please tell us whether you’re going to write to the RAIB and tell them they’ve got it all wrong in their report, as to the cause of this incident?
I’m somewhat unclear as to your point here. You (and @DanNCL) seem to be suggesting that the TPWS was calibrated incorrectly in this case?
Keen to see your evidence.
== Doublepost prevention - post automatically merged: ==
Correct. It performed exactly as it’s supposed to.
Last edited:
Signal Head
Member
- Joined
- 26 May 2013
- Messages
- 466
If you lower the one on approach to the PSR you will get unwanted/unnecessary interventions as it will be too far inside the braking curve. The incident train was well under the set speed and still hit the buffers.
So, your suggestion that the set speed of the existing loops is too high is incorrect.
Your suggestion of a third OSS is sound, and that's essentially how TPWS arrangements are designed, each set of loops with a set speed low enough that the next set can deal with but, are you going to do this everywhere? An OSS on approach to a PSR to ensure the train is inside the braking curve, and another one after the PSR to ensure it's been obeyed? Effectively (in this case) Moorgate Control but with TPWS.
There is a reasonable expectation that drivers will stop trains at terminal stations, unless incapacitated (DSD and Vigilance Device covers that), and in cases of brake failure or adhesion issues TPWS can't help anyway, unless you force the speeds low enough to deal with worst-case adhesion even when it's not.
A friction buffer would have probably helped mitigate the rather serious results here, and I'm a bit surprised one hadn't been provided after previous similar incidents.
DanNCL
Established Member
I'm suggesting that the current TPWS setup at that location isn't sufficient to mitigate against a collision, and thereore isn't fit for purpose. One possible solution would have been to lower the trigger speed of the first OSS.
If it's what's needed to stop more incidents like this that write off entire units and have the potential to injure several people then yes. The UK has one of the safest railway networks in Europe, if not worldwide, but if incidents like this keep happening at the frequency they have been happening over the last few years it risks losing that status. I don't know how many places this happens on the mainline, but I know a version of Moorgate control is facilited by the Tyne & Wear Metro's Indusi system on all of their dead end tracks, not just those underground.
It's a reasonable expectation but as I mentioned in an earlier post TPWS really ought to be designed to take into account that if an application is needed the train likely isn't being driven in the manner that it should be, although obviously that's not always the case for example with worst case adhesion.
niggill617@gma
Member
The OSS are around 300 m from the 15 mph commencement. This covers 10 mph above line speed at 12 percent g, excepting brake force build time. The grids are actually too close. The trip speed could easily be set at 38 mph and fall inline with the TOC policy of step one braking.
Ok then.
As I’ve pointed out several times now, TPWS isn’t intended to mitigate against that kind of incident.
I would suggest that it worked exactly as it’s supposed to, hence why the RAIB didn’t mention it in their report. But what do I know. I only work with TPWS every day!
Are you going to suggest to the powers that be that the current system is inappropriate?!
From the report:
Would revised OSS speeds, and additional OSS loop, or even a friction buffer stops as mention above, not be an appropriate risk control measure.
That's not practical, for reasons mentioned above. If driven normally, the train would have been able to stop fine.
How would you do that?
Well it wasn't driven properly and there was a crash. Should we remove other TPWS loops because when driven normally trains don't activate them?
From recollection, he was messaging about Murray Walker passing away.
Signal Head
Member
- Joined
- 26 May 2013
- Messages
- 466
Trains don't go from 'linespeed' to PSR speed instantly, at the board, they start braking some distance back so that the deceleration is gradual. The further away from the restriction (or signal), the higher the 'permitted' speed will be. The loops need to allow for that, otherwise a driver, driving correctly, will have an unwarranted intervention. The deceleration follows a curve (imagine it drawn on graph paper (speed vs distance), a train that's under the speed expected at a particular point is 'inside' the curve.
In this instance the driver had reduced speed, got past the PSR OSS ok, so was apparently braking for the PSR.
The loops were passed at 42mph (set speed of the loops is 53), so was 'inside' the curve at that point. When it failed to continue to decelerate, at some point it would have crossed over the curve so was then 'outside' it, ie it's unlikely to reduce speed to 15 (in this case) by the PSR board but, there were no further OSS loops to 'detect' the overspeed until the one for the buffer itself.
Where would you draw the line with squashing a TPWS braking curve? 30 mph? 20? 15?
It's designed to mitigate against a driver failing to brake from the normal operating speed of the train in the area. If you want to continuously monitor the speed of the train in that fashion then TPWS isn't what you need, it's something like ERTMS - and that comes at a cost.
TPWS is not a golden bullet and never has been considered as such.
It's designed to mitigate against a driver failing to brake from the normal operating speed of the train in the area. If you want to continuously monitor the speed of the train in that fashion then TPWS isn't what you need, it's something like ERTMS - and that comes at a cost.
TPWS is not a golden bullet and never has been considered as such.
DanNCL
Established Member
If TPWS isn’t able to mitigate against this type of incident then there should be another system that can. Other countries can do it, why should the UK be any different?
You reduce the trigger speed for existing OSS loops at locations where that is feasible, or install an extra OSS loop at locations where modifying an existing OSS loop isn’t appropriate.
Because things change? Paragraph 89 notes all those incidents were before TPWS was fitted.
If you’re prioritising safety investments, it seems reasonable to start with places that have recently had an incident as that implies there is an issue that requires fixing.
The driver didn't brake at all until he put it in emergency on the platform. He accelerated to 50mph leaving Fazakerley then coasted from there. His speed had fallen to 41mph as he entered the platform at Kirkby.
There is another system: BR-ATP, and more recently ETCS, which uses the same principle of continuous speed monitoring. But remember what happened with the fitment of BR-ATP: it was installed on certain lines (GWML/Chiltern) then further rollout was paused as it was too expensive, and TPWS was adopted as a more cost effective alternative. Not as good as ATP, but the cost-benefit calculation (installation costs vs deaths prevented) was deemed to reduce the level of risk to an acceptable level. With an infinite amount of money, I'm sure we'd all like to see ETCS installed across the entire network, but we don't live in a world like that.
bramling
Veteran Member
To be fair, it’s a valid question to ask whether the level of protection provided by TPWS could have been better, and the comparisons with Tyne & Wear (or for that matter LU) are reasonable in my view.
If my memory is right, TPWS was originally intended as a mitigation against SPADs. I might be misremembering, but I seem to recall that terminal protection was included as a “nice to have”, perhaps when someone realised it could be included relatively easily as a spin-off benefit. I’m open to correction on this, though.
It presumably was never intended to deal with a driver operating in this way. An incapacitated driver would likely be picked up either by the DSD, by not acknowledging the AWS, or by the TPWS in the right circumstances. Three different mitigations seems reasonable in my book.
Kirkby does seem to have a fairly specific set of characteristics - a fairly fast final approach, a comparatively short platform, and essentially nothing in the way of an over-run. Walton-on-Naze springs to mind (which, by coincidence, had a fairly similar sort of overrun, and even more coincidentally involving similar traction). Whether it’s worth adding extra speed traps, which will of course likely carry performance disbenefits as well as the monetary costs, is a fair question, though given how rare such collisions are I suspect the answer is it probably isn’t worthwhile making changes.
That said, the very unique setup at Kirkby and Ormskirk with the back-to-back buffer stops probably has room for improvement, especially with Kirkby having a history of such incidents, though is it not the case that the layout there is shortly to change anyway?
Surreytraveller
On Moderation
- Joined
- 21 Oct 2009
- Messages
- 3,905
It also makes one wonder how long the driver has been getting away with it without incident or getting caught
GC class B1
Member
I agree that setting the OSS set speed specifically for the TOC driving policy may have prevented this incident and wouldn’t have a performance penalty as it the lops would only operate for a train that was not being driven in accordance with that policy. The driver seems to have been driving normally until the normal braking point and was then distracted. I understood that TPWS was intended to mitigate the effects of this situation as far as possible. This incident seems to demonstrate that a generic TPWS policy (in accordance with RIS-0775-CCS) is not appropriate for specific risk areas like this.
zwk500
Veteran Member
You've started this by saying your a signalling layman, and now you're declaring the Railway Industry Standard isn't fit for purpose?
TPWS is designed to catch the train if the driver gets it slightly wrong, it isn't designed to prevent incidents when there is a major departure from the anticipated behaviour. Only a PTC system can do so.
What would your solution be to prevent this incident?
edwin_m
Veteran Member
Indeed, it was originally intended to monitor signals and speed restrictions only, either being less demanding than buffer stops. At a signal there is always some distance beyond to the point of any collision, and TPWS aims to stop the train within that distance rather than at the signal itself. At a speed restriction there is some level of safe overspeed. So in both cases there is a bit of leeway between the TPWS supervision speed and the normal braking curve that should be followed, and there is little risk of false trips.
At a buffer stop the train needs to be on a braking curve that goes down to zero at the stop itself, only a few metres beyond the normal stopping place. A relatively crude system such as TPWS is always going to struggle to get that level of control without numerous false activations, as indeed have been seen at buffer stops though some were for other technical reasons to do with accuracy of position detection.
The report mentions that the back-to-back arrangement at Kirkby will disappear when the Merseyrail to Northern transition moves to the new station at Headbolt Lane, and of course the 507/508 units with their rather primitive DSD are also on the way out. If that hadn't been so, it's possible recommendations would have been made to modify both.
- Status