Derailment at Kirkby (Merseyside) - 13/03/2021

[Taunton]

As a signalling layman, I wonder why the TPWS OSS loops didn’t prevent the incident as it is my understanding that TPWS is intended to prevent SPADs (and by extension buffer stop collisions) and overspeeding.

My question likewise. I recall in the cab video of the London Bridge to Brighton charity record run a few years ago, the driver was concerned about approaching the Brighton bufferstops too fast, even when substantially slowed down, and was assured that for the run the TPWS grids had been specially switched off.

 

[zwk500]

My question likewise. I recall in the cab video of the London Bridge to Brighton charity record run a few years ago, the driver was concerned about approaching the Brighton bufferstops too fast, even when substantially slowed down, and was assured that for the run the TPWS grids had been specially switched off.

As @ainsworth74 has quoted, the Driver passed the first OSS at a slow enough speed so the system didn't trip, and by the time he was approaching the 2nd TPWS grid (which had been positioned assuming a defensive approach) he had already dropped the emergency brake so there was nothing more TPWS could do.

 

[43066]

My question likewise. I recall in the cab video of the London Bridge to Brighton charity record run a few years ago, the driver was concerned about approaching the Brighton bufferstops too fast, even when substantially slowed down, and was assured that for the run the TPWS grids had been specially switched off.

This is clearly explained in the paragraphs of the report quoted by @ainsworth74.

Essentially he was slow enough not to trigger the overspeed loops protecting the platform PSR, but too quick for the platform loops to make any difference (+he had the brake in already when they triggered).

 

[DanNCL]

So in other words, even though the driver was at fault, this incident like others would have been prevented if the TPWS loops at that location had been set up with a lower trigger speed.

The purpose of TPWS is to bring a train to a halt before the point where it would collide with anything. Clearly the trigger speed for the TPWS loops at Kirkby is too high, as if they’d been lower the train could have been stopped in time.

 

[43066]

So in other words, even though the driver was at fault, this incident like others would have been prevented if the TPWS loops at that location had been set up with a lower trigger speed.

The purpose of TPWS is to bring a train to a halt before the point where it would collide with anything. Clearly the trigger speed for the TPWS loops at Kirkby is too high, as if they’d been lower the train could have been stopped in time.

Not necessarily. It might well be possible for a train to cross the overspeed loops at around 40mph (as this train did) and still reach a safe stop in the platform if being driven normally. The platform loops will prevent a low speed collision but are not designed to stop a train being driven into the platform at 40mph, so this incident falls between two stools.

 

[GC class B1]

Not necessarily. It might well be possible for a train to cross the overspeed loops at around 40mph (as this train did) and still reach a safe stop in the platform if being driven normally. The platform loops will prevent a low speed collision but are not designed to stop a train being driven into the platform at 40mph, so this incident falls between two stools.

I agree with DanNCL that the OSS set speed is too high especially as the station is not a through platform but in effect a bay platform. Also the speed limit at the start of the platform is 15MPH so surely TPWS should prevent the train being driven into the platform at more than 15MPH unless the driver slows down for the OSS loops and then coasts or accelerates. The platform loops should then prevent a buffer stop collision.

 

[skyhigh]

I agree with DanNCL that the OSS set speed is too high especially as the station is not a through platform but in effect a bay platform. Also the speed limit at the start of the platform is 15MPH so surely TPWS should prevent the train being driven into the platform at more than 15MPH. The platform loops should then prevent a buffer stop collision.

I think that demonstrates a bit of a lack of understanding how TPWS functions.

== Doublepost prevention - post automatically merged: 11 Aug 2022 ==

Edit - I note you have since edited your post

 

[43066]

I agree with DanNCL that the OSS set speed is too high especially as the station is not a through platform but in effect a bay platform. Also the speed limit at the start of the platform is 15MPH so surely TPWS should prevent the train being driven into the platform at more than 15MPH unless the driver slows down for the OSS loops and then coasts or accelerates. The platform loops should then prevent a buffer stop collision.

Well the RAIB don’t agree as “OSS set speed is too high” isn’t mentioned as a causal factor in the report.

I’ve explained above how the incident could have occurred without the loops being incorrectly set. I think you may misunderstand how TPWS works.

 

[NoOnesFool]

Not really:



The cause was the driver's actions (or more precisely their inaction in not applying the brake in time). But while looking at the matter, RAIB have noticed that Merseyrail could be better at dealing with fatigue.

In the investigation in to the Southall rail crash, it was speculated as to whether the driver was distracted by his bag also.

The train stood up well to the impact in this case at Kirkby. It is a good job there were no football fans on the platform (for their sakes).

The driver's actions were irresponsible but a lapse in concentration after a 40 year career with a previously un-blemished record suggests to me that there was no pattern of careless driving. It is sad to see a railwayman's career end this way.

 

[Railsigns]

Many of today's comments in this thread focus on the effectiveness of the TPWS but overlook the vital role that AWS plays. The driver acknowledged three AWS warnings (including two irrelevant ones) on approach to the buffer stop at Kirkby. The role that both systems perform need to be considered together.

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.

 

[Fawkes Cat]

a lapse in concentration

Again, how is choosing to do something other than your job (in this case, sending a message on your phone) a ‘lapse in concentration’?

 

[43066]

The driver's actions were irresponsible but a lapse in concentration after a 40 year career with a previously un-blemisned record suggests to me that there was no pattern of careless driving. It is sad to see a railwayman's career end this way.

It is. Albeit sending WhatsApp messages while driving is monumentally stupid, totally indefensible and would be a sackable offence if discovered, even without a collision happening!

Phone needs to be off and in your bag.

 

[GB]

Why is the platform/track like that anyway? ie why two stops blocks instead of it being joined?

 

[Dai Corner]

Why is the platform/track like that anyway? ie why two stops blocks instead of it being joined?

Off topic, so I'll keep it brief.

Merseyside was carved out of south west Lancashire and decided to modernise and electrify their bit of the Liverpool-Preston line. Lancashire/BR/DfT didn't want to follow suit so through trains no longer ran and the odd arrangements now in place appeared. There are other examples at the extremities of the Merseyrail area I believe.

 

[DanNCL]

Not necessarily. It might well be possible for a train to cross the overspeed loops at around 40mph (as this train did) and still reach a safe stop in the platform if being driven normally. The platform loops will prevent a low speed collision but are not designed to stop a train being driven into the platform at 40mph, so this incident falls between two stools.

And there lies the issue. There’s too much of a reliance on things being done correctly. The vast majority of the time if a train is being driven normally there wouldn’t be a TPWS intervention. TPWS should therefore be set to take into account that there’s a reasonable chance any train that has a TPWS intervention isn’t being driven normally.

Well the RAIB don’t agree as “OSS set speed is too high” isn’t mentioned as a causal factor in the report.

I’ve explained above how the incident could have occurred without the loops being incorrectly set. I think you may misunderstand how TPWS works.

Does it suggest they don’t agree or does it simply suggest they don’t think it’s within their remit, but rather the remit of the RSSB and the ORR?

 

[GB]

Ok thanks. But it has to be said that if they were joined up (even keeping the stops in place) would this, and those other incidents have been quite as bad? I understand you can't account for everything everywhere but it looks like an easy answer for overrun mitigation.

 

[Deafdoggie]

Ok thanks. But it has to be said that if they were joined up (even keeping the stops in place) would this, and those other incidents have been quite as bad? I understand you can't account for everything everywhere but it looks like an easy answer for overrun mitigation.

Only if the other platform is clear. If there were a train in there then I doubt the outcome would be better

 

[Dai Corner]

Ok thanks. But it has to be said that if they were joined up (even keeping the stops in place) would this, and those other incidents have been quite as bad? I understand you can't account for everything everywhere but it looks like an easy answer for overrun mitigation.

What if there had been another train in the platform?

 

[_toommm_]

Ok thanks. But it has to be said that if they were joined up (even keeping the stops in place) would this, and those other incidents have been quite as bad? I understand you can't account for everything everywhere but it looks like an easy answer for overrun mitigation.

You could have then had a train crash into another train. The northern services stop at Kirkby for about twenty minutes, so depending on the time of the crash, it could have led to a Merseyrail EMU crashing into a Northern DMU.

 

[plugwash]

As a signalling layman, I wonder why the TPWS OSS loops didn’t prevent the incident as it is my understanding that TPWS is intended to prevent SPADs (and by extension buffer stop collisions) and overspeeding. The report does not appear to consider whether TPWS could (or should) have prevented the buffer stop collision occurring as a result of the driver being distracted.

TPWS is intended to reduce (but not eliminate) the possibility of collisions. It does this by, among other things, sensing the speed at fixed locations and triggering a brake demand if it is too fast. But it has no idea of the speed between those locations.

The driver, by travelling at significantly less than line-speed essentially fell into a gap in the TPWS protection, he was not going fast enough for the first TPWS to trigger, and he was going too fast for the second TPWS to be effective.

The RIAB did not make any specific recommendations for infrastructure improvements at Kirkby because the station is already due to be converted from it's current double-terminus configuration to a through configuration. They did put in a bunch of reccomendations to review risk-assesment processes. though.

Presumablly if it was deemed appropriate to improve the TPWS protection, a third set of loops would be needed.

 

[43066]

And there lies the issue. There’s too much of a reliance on things being done correctly. The vast majority of the time if a train is being driven normally there wouldn’t be a TPWS intervention. TPWS should therefore be set to take into account that there’s a reasonable chance any train that has a TPWS intervention isn’t being driven normally.

I agree. In this case a train being driven perfectly normally could have crossed the PSR OSS at the speed this train did, well below the trigger speed, hence no intervention was triggered.

The entire railway safety regime is predicated on people doing their jobs correctly and not making errors/committing negligent acts such as those that occurred in this case.

Does it suggest they don’t agree or does it simply suggest they don’t think it’s within their remit, but rather the remit of the RSSB and the ORR?

If a safety system malfunctioned/was deliberately isolated and that contributed to an incident it absolutely would be part of their remit, yes. In this case the system worked precisely as intended. TPWS doesn’t provide constant monitoring/oversight of train speed/driver actions and it is not a silver bullet.

 

[ainsworth74]

The driver's actions were irresponsible but a lapse in concentration after a 40 year career with a previously un-blemished record suggests to me that there was no pattern of careless driving.

I'm not entirely sure it was unblemished...

25 Merseyrail provided RAIB with the driver’s employment record, which showed that he had been involved in three safety and six operational incidents between 1995 and 2019. The last two of these occurred within a six-month period in 2019, which triggered Merseyrail’s ‘Individual Development Programme’ (IDP) process. Merseyrail placed the driver on an IDP, which included a post-incident in-cab assessment of the driver’s behaviour, and the analysis of two OTDR downloads to check his driving techniques. The driver was signed off as having successfully completed this process on 20 October 2020.

It is sad to see a railwayman's career end this way.

I do agree on that however.

 

[bramling]

So in other words, even though the driver was at fault, this incident like others would have been prevented if the TPWS loops at that location had been set up with a lower trigger speed.

The purpose of TPWS is to bring a train to a halt before the point where it would collide with anything. Clearly the trigger speed for the TPWS loops at Kirkby is too high, as if they’d been lower the train could have been stopped in time.

I don't think this quite follows. TPWS monitors speed at a particular location. It isn't a continuous ATP system which monitors braking curves.

Obviously TPWS *could* be configured to do this, albeit crudely, by adding many more loops on the approach to a terminus in gradually decreasing steps. But to do so, it would have to be deemed worthwhile, which presumably hasn't been the case.

 

[43066]

I do agree on that however.

I could have understood it (although still not defended it) a little more if he’d had some major life crisis happening in the background. He was texting about something really inconsequential IIRC? I can’t remember where that was reported.

 

[bramling]

In the investigation in to the Southall rail crash, it was speculated as to whether the driver was distracted by his bag also.

I'm going to be a little controversial here, however in the case of Kirkby I'm slightly sceptical about this whole thing over the driver's bag. My rationale for being sceptical is the driver has just driven into a buffer stop, some form of reason had to be offered for the incident happening, and one which wasn't "I was using my phone". It's a little too text-book for my slightly suspicious mind.

 

[ainsworth74]

I could have understood it (although still not defended it) a little more if he’d had some major life crisis happening in the background. He was texting about something really inconsequential IIRC? I can’t remember where that was reported.

Absolutely. It's why I had a lot of sympathy for that driver that derailed their 66 by running off the end of a siding and then had a 170 clip it near Bromsgrove. The reason they were distracted by their phone leading up was worrying about how they'd manage their childcare having just had a national lockdown announced hours before! Very understandable that your mind might not fully be on the job. Shouldn't have done it of course but I can at least understand it.

This incident? I'm not sure the detail was ever published but like you I seem to recall that it was reported somewhere as being trivial.

 

[plugwash]

96 One factor in the process was to consider whether there was any history of buffer
stop collisions at the location within the last five years. This timeframe meant that
the three previous accidents at Kirkby (paragraph 89) were not considered and
did not have any impact on the resulting FWI score. If they had, it is possible that
improvements may have been made at Kirkby to better control the risk of a train
overrunning the buffer stop.

This to me to be one of the most damning paragraphs in the report, if you are going to consider previous incidents why the hell would you restrict yourself to a window as short as 5 years?!

 

[DelW]

Our old friend fatigue again.

Not really:

Indeed, since paragraph 101 states that there was no evidence that fatigue was a causal factor.

 

[6Gman]

I'm going to be a little controversial here, however in the case of Kirkby I'm slightly sceptical about this whole thing over the driver's bag. My rationale for being sceptical is the driver has just driven into a buffer stop, some form of reason had to be offered for the incident happening, and one which wasn't "I was using my phone". It's a little too text-book for my slightly suspicious mind.

I must confess I thought something similar.

Unlike the phone use, where evidence could be secured, there is only the driver's word on the fallen bag and rolling drink bottle. Perhaps it happened; perhaps it didn't. Only one person knows.

 

[Signal Head]

So in other words, even though the driver was at fault, this incident like others would have been prevented if the TPWS loops at that location had been set up with a lower trigger speed.

The purpose of TPWS is to bring a train to a halt before the point where it would collide with anything. Clearly the trigger speed for the TPWS loops at Kirkby is too high, as if they’d been lower the train could have been stopped in time.

Which OSS are you going to set at a lower speed?

The one in the platform is set at 12.5 (standard setting, increased from the original 10 after some 'false' trips with trains at borderline speed).

The one on approach to the PSR is set for the speed allowed according to the braking curve.

What would you change, and how would it work?