Lumo overspeed incident at Peterborough (17/04/2022)

[edwin_m]

Aircraft pilots are also expected to be able to take their vehicle off the planned route with little-to-no notice, and be able to safely get back on the land even with zero outside communication.

Aircraft don't have to interact with fixed infrastructure in the way that railways do. Flying from A to B is pretty similar no matter where A and B are in the world. They also have the fallback that if two planes find themselves on conflicting routes, they would normally be alerted by ATC or by the TCAS system and crucially the pilot can change course to avoid collision at any time when so alerted or simply if seeing another aircraft on a conflicting trajectory. By contrast, if two trains find themselves on the same section of track, the risk that they would then collide is much greater, so there has to be much more emphasis on preventing that happening in the first place.

The details about individual airports are available in written form as already mentioned, and although their layouts can be complex, taxiing is low speed and the pilot can stop the aircraft immediately if lost or if they are heading for collision with another aircraft. Trains pass through complex layouts at speeds that make it impossible to stop if an unsafe situation develops, as happened with this incident.

 

[berwicksfinest]

As the report mentions somewhere (only had chance to skim read most of it), part of the problem here is the combined distance from release point to signal, and signal to turnout. Current standards provide stipulations for both of those separately but don't, so far as I am aware, consider the combination.

In this instance, whilst everything is compliant, I cannot help but feel that, notwithstanding any failures in training, 'we' (ie the S&T function) have to some extent 'set a trap' for drivers to fall into - if the JI is not observed or does not register properly.

I do not subscribe to the view that these things are simply down to driver error all the time, or that they should be written off as such. To do that risks potentially overlooking a 'system failure' (not necessarily a technical one), which has been exposed by a momentary human lapse. That then presents an opportunity to do something about it, to close the loophole or whatever.

The fact that two drivers have almost come to grief here only a year apart tells me that, however compliant it is on paper, the signalling here would probably benefit from another coat of looking at.

As someone who does sit in the big boys chair and is based at Peterborough. I have never seen or hear of problems coming off the Up main into Platform 1.

HOWEVER what I am seeing a lot more of is Train companies nudging Drivers more to complete route learning faster, remember route learning is non productive. and TOCs, and FOCs hate non productive, So a lot more "are you ready to sign yet" is heard.
Herein lies the problem, any body with any experience will say "you don't get pressured into signing something your not ready to" BUT experience gives you the tools to tell your Line Manager to "Foxtrot Oscar" you need more time.
But a lot of Street to Seat or just been made up, really don't want to rock the boat so unfortunately sign it.

The signalling hardware is still the same, yet both these incidents happened recently.

 

[Aviator88]

There isn't any "just" about this. The TD feed from sigbalboxes to sites like traksy is wildly off levels of reliability that you could use for safety. Adapting that to exchange data via WiFi would add a whole new layer of complexity and failure.

In fact building this level of safe comms has been happening for a number of years, and would happen in ECTS, which is essentially ready now and being rolled out much more quickly than your suggestion, and provides a much higher level of safety.

Another factor with maps is simply getting the map to clearly show things like speed differentials. Look at the maps into major stations like Peterborough and tell me you can at a glance figure out which routes are low speed.

This is why I used the words 'or similar', to include ballaises (as used in ETCS) et al.

ETCS provides the 'what', but not the 'why' - yes, route knowledge should be sufficient to understand why you're being checked down to, say, 30 mph, but I think there's definitely an argument for being able to increase the information provided to the driver whilst still avoiding overload.

No need to display speeds on the maps - train drivers (should) know their route like the back of their hands, as they should the route maps themselves. My suggestion isn't to provide reliance on the data, but for it to be used to augment all the other tools and feeds available to the driver.

 

[Railsigns]

Bear in mind that any new system that provides train drivers with current information about the route has to be kept up to date or else it becomes a danger.

Whenever changes are made to the infrastructure, projects will be burdened with the cost of updating these systems at a time when the rail industry is under pressure to reduce its costs.

 

[66701GBRF]

Costs aside, what would be the pro's and con's, and is it technically possible, to have the JI flash...at least at some high risk locations?

 

[driver9000]

As someone who does sit in the big boys chair and is based at Peterborough. I have never seen or hear of problems coming off the Up main into Platform 1.

HOWEVER what I am seeing a lot more of is Train companies nudging Drivers more to complete route learning faster, remember route learning is non productive. and TOCs, and FOCs hate non productive, So a lot more "are you ready to sign yet" is heard.
Herein lies the problem, any body with any experience will say "you don't get pressured into signing something your not ready to" BUT experience gives you the tools to tell your Line Manager to "Foxtrot Oscar" you need more time.
But a lot of Street to Seat or just been made up, really don't want to rock the boat so unfortunately sign it.

This is a very important point and should ram home the importance of thorough route knowledge and assessment of that knowledge. Taking the full allocated training days/trips/hours and extra if necessary is to your benefit because it's the Driver who ends up answering questions if things go wrong. One of the stand out paragraphs in the report to me was the lower than standard number of handling hours the Driver had attained before being passed as competent. They had less than 200 hours with an Instructor but another operator on the same route has 300 hours as their standard. This hasn't done a trainee any favours and I would hope Lumo has increased it's trainee handling hours and reassessed its route pass out questioning since this happened.

I don't sign Peterborough so can't comment on the location specifically but I do train new Drivers to drive long distance express trains and we aim for nearly 300 hours. I also impress on them the importance of fully understanding what you're looking at when reading signals especially at locations where there is the risk of doing something out of the ordinary.

 

[Peter Mugridge]

As someone who does sit in the big boys chair and is based at Peterborough. I have never seen or hear of problems coming off the Up main into Platform 1.

Isn't the problem that usually when a train is going into platform 1 it will be stopping there rather than just passing through, so normally would be slowing down well before the points in question?

 

[harz99]

Just a thought, iirc the fast to slow move for Huntingdon can be taken at a much higher speed than the route via P1 at Peterborough, is there any reason why the passing manoeuvre could not be done there rather than PBO, given that both LUMO and LNER are running at full line speed the following train isn't going to be held up?

 

[berwicksfinest]

Isn't the problem that usually when a train is going into platform 1 it will be stopping there rather than just passing through, so normally would be slowing down well before the points in question?

Not necessarily as coming through Platform 1 you can be dropped up to Fletton to allow something to pass on the Main, Grand Central do exactly that quite regularly. and ordinarily anything London bound from the North uses platform 3, Platform 1 generally used for the Horsham's and EMR Lincoln's

Just a thought, iirc the fast to slow move for Huntingdon can be taken at a much higher speed than the route via P1 at Peterborough, is there any reason why the passing manoeuvre could not be done there rather than PBO, given that both LUMO and LNER are running at full line speed the following train isn't going to be held up?

Up main in at Huntingdon is a 40 mph set of points again brought in with a flashing 4 aspect countdown from Stukeley bank

 

[HSTEd]

This is why I used the words 'or similar', to include ballaises (as used in ETCS) et al.

ETCS provides the 'what', but not the 'why' - yes, route knowledge should be sufficient to understand why you're being checked down to, say, 30 mph, but I think there's definitely an argument for being able to increase the information provided to the driver whilst still avoiding overload.

Ultimately, is it critical to know "why" you are being checked down in speed?

If the train goes over the junction at a safe speed then does it really matter very much?

 

[TheBigD]

Not necessarily as coming through Platform 1 you can be dropped up to Fletton to allow something to pass on the Main, Grand Central do exactly that quite regularly. and ordinarily anything London bound from the North uses platform 3, Platform 1 generally used for the Horsham's and EMR Lincoln's

This train is booked to be overtaken at Peterborough every weekday. It is booked to be routed up fast to Platform 1.

Realtime Trains | 1A64 0954 Sunderland to London Kings Cross | 27/07/2023

Real-time train running information for 1A64 0954 departure from Sunderland to London Kings Cross on 27/07/2023. From Realtime Trains, an independent source of train running info for Great Britain.

www.realtimetrains.co.uk

 

[Signal Head]

I'm surprised that, in this day and age, modern trains don't have a display in the cab which shows the local track diagram, as well as what route has been set by the signaller in real-time etc. I know the point is to keep the driver's eyes outside, but it would assist the driver in moments of doubt without having to spend time contacting the signaller.

There's no need to rely on GPS - just use the information collected from the track circuits and relay it via Wi-Fi on the track infrastructure or similar.

The main problem with your post is your use of the word 'just'.

 

[800301]

While I know it’s a now somewhat outdated system that isn’t fitted to this line, I believe ATP (Automatic Train Protection) if fitted would have prevented this, as despite the signal stepping up ATP would give you a release speed that you can’t exceed until you have passed the signal, not only that but it would also know the speed of the diverging route which would be displayed to the driver, obviously ETCS would as well.

GPS systems have proven many times to not be reliable on the railway and that’s just dealing with doors

 

[david1212]

ECTS is coming and when operational would I presume adjust the speed limit approaching the points according to how they are set so prevent any incident like this.

Looking back to when either the current signalling system was installed of if later when TWPS was added as was mentioned at least once several pages back is had a TWPS transmitter grid been installed which was only enabled if the points were set for the platform in both this and the previous instance the train would have been slowed if not to the speed limit at least to a significantly lower speed. Can this be implemented? Hindsight is a wonderful thing but if it can a question is why it was not done. There must be numerous similar configurations around the network.

 

[Aviator88]

Ultimately, is it critical to know "why" you are being checked down in speed?

If the train goes over the junction at a safe speed then does it really matter very much?

It's not safety critical, just a concept of increased spacial awareness = better, especially if/when the technology is there.

The main problem with your post is your use of the word 'just'.

Apologies, I come from an industry that isn't chronically underfunded

 

[JN114]

While I know it’s a now somewhat outdated system that isn’t fitted to this line, I believe ATP (Automatic Train Protection) if fitted would have prevented this, as despite the signal stepping up ATP would give you a release speed that you can’t exceed until you have passed the signal, not only that but it would also know the speed of the diverging route which would be displayed to the driver, obviously ETCS would as well.

Release speed wouldn’t help here, that’s just for crawling up to red aspects and stopping before conflict if you SPAD. With such a long distance to conflict point (700+ metres) I expect release speed on this signal would be right up at the maximum (so 50mph).

As for limiting speed over the junction, it would depend on the infrastructure fit. If it was comprehensive like Chilterns, or GWML east of Airport Jn then yes, after passing the junction signal you’d get a target speed of 30 at the indication point for the divergence. But if it were a fast lines only fit, like the GWML Mains-only fit west of Airport Jn; I don’t think you get a target speed once you pass the last signal beacon before exiting, it just reverts to maximum vehicle speed supervision (so 125 in this case).

 

[AlexNL]

I'm surprised that, in this day and age, modern trains don't have a display in the cab which shows the local track diagram, as well as what route has been set by the signaller in real-time etc. I know the point is to keep the driver's eyes outside, but it would assist the driver in moments of doubt without having to spend time contacting the signaller.

There's no need to rely on GPS - just use the information collected from the track circuits and relay it via Wi-Fi on the track infrastructure or similar.

The Dutch railways (NS) have a system somewhat like this. There's an app which NS drivers run on their tablet (on some trains it's even integrated), and it gives them information about the state of the track ahead (and a bit behind them) - showing what route a train will take and also what actions other trains will be doing (seen relative to the driver).

You can see it in action in this video (leftmost screen in the cab), if you look closer towards the end of the video you can see the conflicting moves get mentioned:

 

[Mattydo]

Aircraft use GPS augmented with IRS and radio nav updates for long range en route navigation. Terminal navigation of the accuracy needed to conduct an approach using GPS is (bizarrely still) in its infancy and the IRS (and consequently radio beacon) gross error checks are made more often to ensure accuracy or ground equipment is used to negate it. Whilst the actual accuracy often exceeds the minimum required guaranteed accuracy is still in the order of metres. Not sufficient to necessarily track exactly which track a train is on.

GPS is incredibly accurate but it is deliberately degraded for non military use and subject to the potential for further blocking at times.

Long winded way of saying GPS probably isn't the answer here without a lot of extra kit onboard to compensate that would probably take as much time to develop test and approve than existing or incoming tech.

TPWS OS loops would seem to be the obvious answer but that's not necessarily a straightforward solution either, when ETCS is on the way with in cab route information and speed limitation.

In the interim one would hope further highlighting the deficiency in route knowledge or risk awareness would suffice. It works in a lot of similar areas.

 

[Sly Old Fox]

Release speed wouldn’t help here, that’s just for crawling up to red aspects and stopping before conflict if you SPAD. With such a long distance to conflict point (700+ metres) I expect release speed on this signal would be right up at the maximum (so 50mph).

As for limiting speed over the junction, it would depend on the infrastructure fit. If it was comprehensive like Chilterns, or GWML east of Airport Jn then yes, after passing the junction signal you’d get a target speed of 30 at the indication point for the divergence. But if it were a fast lines only fit, like the GWML Mains-only fit west of Airport Jn; I don’t think you get a target speed once you pass the last signal beacon before exiting, it just reverts to maximum vehicle speed supervision (so 125 in this case).

That’s not true, ATP will give you the speed limit of any crossover from main to relief, even where the reliefs aren’t fitted.

In this case you’d get a flashing target speed of 30mph for the junction once the signal protecting it was showing a proceed aspect. It will then revert to ‘off’/partial supervision mode as the train traverses the junction.

 

[jupiter]

Many years back in a former life I worked on timing systems for telecommunications networks. GPS has a very, very accurate one second rising edge signal that would be ideal for synchronising exchanges but i) you’d have a put a hole in every building for an aerial and ii) the system is outside your control.

This has happened twice in a fairly short space of time now? The first thing any good lawyer for injured parties does is find out if there is any history, and if so what was, or could have been, done to prevent it happening again. I’d say a stance of “further driver education” and “wait for six years for a new system” is indefensible.

I appreciate signalling system changes are incredibly complex and expensive, but you’re looking at the possibility of a very serious accident here.

 

[takno]

Many years back in a former life I worked on timing systems for telecommunications networks. GPS has a very, very accurate one second rising edge signal that would be ideal for synchronising exchanges but i) you’d have a put a hole in every building for an aerial and ii) the system is outside your control.

This has happened twice in a fairly short space of time now? The first thing any good lawyer for injured parties does is find out if there is any history, and if so what was, or could have been, done to prevent it happening again. I’d say a stance of “further driver education” and “wait for six years for a new system” is indefensible.

I appreciate signalling system changes are incredibly complex and expensive, but you’re looking at the possibility of a very serious accident here.

You're looking at a very low probability of a serious accident, as evidenced by the fact that in the first case the train took the points at 3x the speed limit and nothing very significant actually happened, and in the second case it's not clear that any passengers even noticed. Out of an excess of caution which isn't shown in most industries, these near-misses were recorded and investigated by a very professional independent body of experts for a year. With all due respect a couple of chippy lawyers kicking up a fuss isn't going to lead to a more useful result for anybody

 

[Bill57p9]

Yes, Drivers are issued (or should be) with maps for route learning guidance. I keep certain sections in my bag mainly for reference to shunting moves at large stations that may not happen all that often but a Driver likely wouldn't have them out as a matter of routine.

Thanks @driver9000 .
Presumably there is a process in place to ensure driver’s charts stay up to date. (In UK aviation the AIP is the single point of truth, supplemented by NOTAMs for urgent and/or temporary amendments. The challenge is ensuring that everyone uses the current version. Internationally aviation has a 4 weekly change cycle called an AIRAC).

 

[Ediswan]

GPS is incredibly accurate but it is deliberately degraded for non military use and subject to the potential for further blocking at times.

That practice ended in 2000. Where military GPS has an advantage is the use of dual frequency receivers. But these can now be found in some Apple devices.
https://www.gps.gov/systems/gps/performance/accuracy/

Is military GPS more accurate than civilian GPS?​

The user range error (URE) of the GPS signals in space is actually the same for the civilian and military GPS services. However, most of today's civilian devices use only one GPS frequency, while military receivers use two.
Using two GPS frequencies improves accuracy by correcting signal distortions caused by Earth's atmosphere. Dual-frequency GPS equipment is commercially available for civilian use, but its cost and size has limited it to professional applications.
With augmentation systems, civilian users can actually receive better GPS accuracy than the military. Learn more

 

[MotCO]

I'm conscious that some of the posts on this thread are moving into 'speculative discussion' territory rather than discussing the original incident, and I'm as guilty as anyone! I have therefore set up a separate thread to look at suggestions for avoiding such incidents in the future here https://www.railforums.co.uk/thread...-safer-to-avoid-another-lumo-incident.251559/

 

[Dai Corner]

That practice ended in 2000. Where military GPS has an advantage is the use of dual frequency receivers. But these can now be found in some Apple devices.
https://www.gps.gov/systems/gps/performance/accuracy/

This bit from the same website would be enough to put me off using GPS for anything safety-critical

Why does GPS sometimes show me in the wrong place?
Many things can degrade GPS positioning accuracy. Common causes include:

Satellite signal blockage due to buildings, bridges, trees, etc.
Indoor or underground use
Signals reflected off buildings or walls ("multipath")

 

[zwk500]

Thanks @driver9000 .
Presumably there is a process in place to ensure driver’s charts stay up to date. (In UK aviation the AIP is the single point of truth, supplemented by NOTAMs for urgent and/or temporary amendments. The challenge is ensuring that everyone uses the current version. Internationally aviation has a 4 weekly change cycle called an AIRAC).

Changes to the Sectional Appendix are circulated in Weekly Operating Notices. The Sectional Appendix shows the permitted linespeed for all lines and crossovers.

 

[driver9000]

Thanks @driver9000 .
Presumably there is a process in place to ensure driver’s charts stay up to date. (In UK aviation the AIP is the single point of truth, supplemented by NOTAMs for urgent and/or temporary amendments. The challenge is ensuring that everyone uses the current version. Internationally aviation has a 4 weekly change cycle called an AIRAC).

The Sectional Appendix is updated via the PON (Periodical Operating Notice) every 3 months as necessary but the Weekly Operating Notice is be used between issues of the PON for alterations. The SA doesn't carry such detail of signal locations and routings so Drivers should be issued with more detailed maps which can be used for route learning guidance hopefully accompanied by a routing index for junction signals. These are either produced in house by the TOC training department or an Instructor or the other option is to the maps produced by Track Access Productions. The level of detail varies and the most detailed maps I've used were made by First North Western which went to the detail of naming the roads carried by bridges!

 

[Aviator88]

The Dutch railways (NS) have a system somewhat like this. There's an app which NS drivers run on their tablet (on some trains it's even integrated), and it gives them information about the state of the track ahead (and a bit behind them) - showing what route a train will take and also what actions other trains will be doing (seen relative to the driver).

You can see it in action in this video (leftmost screen in the cab), if you look closer towards the end of the video you can see the conflicting moves get mentioned:

Cheers for that, very interesting to see!

I think, regarding the argument that considers a driver may become over-reliant on such a system, it's important to bear in mind that new drivers joining the industry (in the UK at least) sit a small battery of tests that include their ability to concentrate and prioritise tasks such as these.

Coming from an aviation background, I've always found that increasing technology has only made my job easier, and it's very easy to learn to filter out less important tasks in a variety of situations.

 

[Mattydo]

That practice ended in 2000. Where military GPS has an advantage is the use of dual frequency receivers. But these can now be found in some Apple devices.
https://www.gps.gov/systems/gps/performance/accuracy/

Sorry I should have been clearer... Can be. Anybody flying in the eastern Ankara FIR in the last decade can attest to that. Unlikely in the UK but I'm not sure a purely GPS system would pass muster as a safety system whilst the possibility, however remote, remains. Point is the system needed (ETCS) already exists and is approved and anything else would have to go through research and testing and probably take longer.

Anyway as stated above this is all hugely speculative.

 

[Signal Head]

Many years back in a former life I worked on timing systems for telecommunications networks. GPS has a very, very accurate one second rising edge signal that would be ideal for synchronising exchanges but i) you’d have a put a hole in every building for an aerial and ii) the system is outside your control.

This has happened twice in a fairly short space of time now? The first thing any good lawyer for injured parties does is find out if there is any history, and if so what was, or could have been, done to prevent it happening again. I’d say a stance of “further driver education” and “wait for six years for a new system” is indefensible.

I appreciate signalling system changes are incredibly complex and expensive, but you’re looking at the possibility of a very serious accident here.

They aren't - always. If this were in a relay interlocking it could be a relatively simple issue to alter the controls.

Once the systems are computer based, you're into data changes. In theory, that should be 'less work' and therefore easier/cheaper - however, there are fewer companies able to undertake the work, and in some instances, depending on the systems used, the alterations can only be undertaken by the supplier of the kit owing, in part, to software rights issues I believe..

I'm personally aware of two instances - in one, the quote for the data change to update the train describer layout map was around double the price for the design alterations to the (relay) interlocking.

In the other, which involved some minor changes to the VDU layout at the control centre, the data changes were a staggering 10x the price of the interlocking design change. Had the interlocking still been worked by the original panel, it would have been a simple matter of a few new panel tiles, or even a 'DIY' job with some Humbrol paints!