Croydon Tram Crash

[bramling]

Let me ask this question then, if the drivers route knowledge tells him that the corner is there, then why has there been possibly previously two non fatal incidents of speeding on that bend, before the most recent fatal incident? Unless all three incidents are down to drivers blacking out, I would say that there is a need for some drivers to get a better idea of what point to be slowing down before the corner in question and as such place warning speed signs plus safety braking equipment like TPWS prior to the car to make sure that the driver is putting on the brakes of the tram in time and not just at the end of the tunnel, just prior to be going round the bend. That way you can make sure as you state. that the the curve is driven appropriately.

There are numerous reasons why a driver might not reduce speed in time for a speed restriction.

First of all, the driver could brake in the normal place but be caught out by low rail adhesion. Since this incident I've been surprised to learn that the hazard brake on trams can be more badly affected by poor rail conditions than I previously thought. It's worth noting from photos that the immediate approach to the curve from the Lloyd Park direction seems to be very prone to leaf-fall, and bear in mind the morning concerned was also raining. Nothing has been released officially which confirms or denies whether poor rail conditions may have played a part. Obviously it's part of a driver's skill to anticipate poor rail conditions, but humans make errors of judgement ...

Then the driver could be incapacitated by some kind of medical issue. Remember this has been postulated as a possible cause of the Moorgate accident. History and experience shows a deadman device can't always be relied upon to be effective.

You can also add fatigue to the last possibility.

Then some kind of distraction, either a device like a mobile phone, or even something simple like reading a book (personally I find the latter unbelievable, but where I am drivers openly talk about reading books or newspapers whilst driving trains so never say it doesn't happen), even looking at a duty book at an unfortunate moment, perhaps distracted by something going on inside the tram, something unusual going on outside, or a personal issue weighing on the mind. All it takes is the mind to veer off-task for a few seconds and an error can happen.

Finally it's possible to simply have a good old-fashioned lapse of concentration. Anyone who performs a repetitive task (and driving a tram on a fairly limited selection of routes, stopping at every station and where all of the stations look pretty much the same must rank pretty highly for boredom) will tell you that it's quite possible to lose concentration for seemingly no reason. There is a particular "trap" on the railway where a driver can see a yellow or red signal, one which they normally see showing that colour, then have a lapse of concentration and make no further reaction until the last minute. This has been a factor in many SPADs which I have investigated. One can only conclude that the mind normalises a situation where a reaction (i.e. braking) is normally/always required, and allows itself to switch off, without actually making the correct reaction. If you add in the possibility of confusing one's location, again quite likely to happen when performing a repetitive task, and it's again easy to see how errors can occur.

The above list is in no way exhaustive.

You can have all the warning signs in the world, but it doesn't necessarily prevent any of the above from occurring.

Whilst it's good to know the root cause of an incident, when it comes to devising preventive measures it can be irrelevant -- you simply devise a mitigation which covers all the possibilities you possibly can, within the budgetary and technical constraints imposed upon you. This is what was done after Moorgate, and I suspect is what will happen here.

 

[littlerock]

interesting account of presentation of initial report to H of C

http://www.croydonadvertiser.co.uk/...-explains-mp/story-29894756-detail/story.html

esponding to shadow minister for London Andrew Slaughter's question as to whether this "specific" matter could be raised as part of the three ongoing investigations into the crash, Croydon Central MP Gavin Barwell said "lots of constituents have already raise that issue" with him.

And he explained: "Essentially, trams are buses on rails. Because they run part of their route on rails and part of it on roads, trams have to rely, at least for part of the route, on drivers driving according to the conditions in front of them.

"Therefore, trams cannot have the same kind of signalling systems as trains.

"However, there is a legitimate question about sections of the route where trams run on rails and are akin to trains, and I am sure that that will be one of the issues addressed in the investigation."

This raises the prospect that parts of Croydon's tram network could become automated depending on the outcome of the investigations....

worth a read as it suggests regs might need to be modified where the tram is effectively operating as a train and automatic braking might be needed on the "train "type tracks.

 

[Antman]

First of all, the driver could brake in the normal place but be caught out by low rail adhesion. Since this incident I've been surprised to learn that the hazard brake on trams can be more badly affected by poor rail conditions than I previously thought. It's worth noting from photos that the immediate approach to the curve from the Lloyd Park direction seems to be very prone to leaf-fall, and bear in mind the morning concerned was also raining. Nothing has been released officially which confirms or denies whether poor rail conditions may have played a part. Obviously it's part of a driver's skill to anticipate poor rail conditions, but humans make errors of judgement ....

I suspect that this is the reason but obviously we'll have to wait and see.
--- old post above --- --- new post below ---

interesting account of presentation of initial report to H of C

http://www.croydonadvertiser.co.uk/...-explains-mp/story-29894756-detail/story.html

worth a read as it suggests regs might need to be modified where the tram is effectively operating as a train and automatic braking might be needed on the "train "type tracks.

Presumably the off street sections could be fully automated with the driver still in situ? Whether the costs could be justified is another matter.

 

[AlterEgo]

There is a focus throughout on line of sight and the assumption is pretty much made that this is always available

When is line-of-sight not available? It is a concept relating to the operation of the tram, not a description of a clear day on straight track for example.

 

[Domh245]

Do we know what data this contains for the Croydon tram?

At the moment, only what the RAIB has told us, which is the speed at which it came off, and the fact that there was some kind of braking activity in the 180m "braking zone" before the speed restriction sign. When the full RAIB report comes out, we'll see a lot more.

 

[AlterEgo]

At the moment, only what the RAIB has told us, which is the speed at which it came off, and the fact that there was some kind of braking activity in the 180m "braking zone" before the speed restriction sign. When the full RAIB report comes out, we'll see a lot more.

And to manage expectations, bearing in mind this accident claimed seven lives and resulted in the arrest of the driver on suspicion of causing those deaths, it might take a year, or more.

James Street's RAIB report took a year to be produced and occurred alongside a ME investigation and the prosecution of the guard.

 

[mirodo]

I suspect that this is the reason but obviously we'll have to wait and see.

The point at which the team should have commenced braking is in tunnel and therefore free of leaf fall.

 

[Antman]

The point at which the team should have commenced braking is in tunnel and therefore free of leaf fall.

Assuming you meant the tram, the area between the tunnel and the curve is the likely problem. Even in the tunnel the rails are likely to be damp.

 

[bramling]

Presumably the off street sections could be fully automated with the driver still in situ? Whether the costs could be justified is another matter.

I think it would be very difficult, as (unlike the older parts of Metrolink) the off-street sections in Croydon are quite open. You'd need to completely fence them off, and design out any crossings etc.

It could be possible to install a continuous ATP-type system which simply monitors speed but does not monitor limit of movement authority, although as a bonus you could perhaps include protection for traffic signals etc.

The alternative is some kind of trainstop system. I suspect this may prove the most cost-effective option.
--- old post above --- --- new post below ---

Assuming you meant the tram, the area between the tunnel and the curve is the likely problem. Even in the tunnel the rails are likely to be damp.

Looking at Google Earth, it seems the section where the tram should have been slowing down is about half in tunnel and half in the open.

It's not impossible for poor adhesion to extend into a tunnel, although I'd say it's rather less likely.

Of course poor rail conditions certainly wouldn't have helped *if* the driver braked late, for whatever reason.

 

[Johnuk123]

Re-opening this afternoon.

 

[mirodo]

Assuming you meant the tram, the area between the tunnel and the curve is the likely problem. Even in the tunnel the rails are likely to be damp.

Yes, for some reason my phone keeps autocorrecting tram to team. My point was that although poor rail adhesion may have exacerbated the problem by failing to allow a late braking manoeuvre, the tram should have been braking earlier.

 

[BestWestern]

The problem of course if it was poor rail conditions, is that none of these systems would be able to do anything more than apply the brakes - which a driver would already have done. In that scenario, perhaps something like a sander device would be a more beneficial innovation. I presume these are not already used on trams?

 

[rebmcr]

What I could not find, but am happy to be corrected, were any specific recommendations on what precautions were needed to avoid accidents in areas where a tram is unsighted, ie coming up suddenly to a sharp bend or corner beyond which line of sight is not available in a location where it is isolated from other traffic and pedestrians ,(as at Sandilands). And also where this occurs before or after a tunnel . There seemed a tacit assumption that line of sight is always available on a tram.

If the tram is coming up to a blind corner, the line of sight is indeed reduced to a much shorter distance than in other locations.

The tram driver is expected to slow down to the degree of being able to stop if he suddenly sees a shopping trolley on the tracks around the corner (for example) -- essentially the tram is supposed to be driven as though the track that remains unseen just doesn't exist.

This is in contrast to the mainline railway, where a driver is provided with a guarantee (via the signalling systems) that a specific stretch of track ahead is definitely clear, even if it it not yet in sight. This is known as Movement Authority.
--- old post above --- --- new post below ---

Since this incident I've been surprised to learn that the hazard brake on trams can be more badly affected by poor rail conditions than I previously thought.

I also believed they had such a strong adhesion that leaf mulch would not affect them significantly.

Given that they use an electromagnetic force to clamp themselves to the rail, instead of relying on gravity as the wheels do: would it be possible that even when adhesion is low and they cannot prevent sliding through sideways friction, that they would still exert enough downward EM force to prevent a derailment? Has there ever been research into that possible effect?

 

[Domh245]

The problem of course if it was poor rail conditions, is that none of these systems would be able to do anything more than apply the brakes - which a driver would already have done. In that scenario, perhaps something like a sander device would be a more beneficial innovation. I presume these are not already used on trams?

Sanders are used on Tramlink, and I think that they automatically deploy sand if the tram is experiencing wheelslip/slide. It is possible of course that the tram went out without enough sand in the hoppers.

 

[bramling]

The problem of course if it was poor rail conditions, is that none of these systems would be able to do anything more than apply the brakes - which a driver would already have done. In that scenario, perhaps something like a sander device would be a more beneficial innovation. I presume these are not already used on trams?

No idea about whether trams have sanding gear, however regarding speed monitoring obviously some form of overlap would have to be built in to allow for poor rail conditions. A sophisticated system could calculate a braking curve, whilst a more crude system would have to have speed reduced in steps. One way or other, the start of the 12 mph would probably have to be moved further towards Lloyd Park.

 

[edwin_m]

No idea about whether trams have sanding gear, however regarding speed monitoring obviously some form of overlap would have to be built in to allow for poor rail conditions. A sophisticated system could calculate a braking curve, whilst a more crude system would have to have speed reduced in steps. One way or other, the start of the 12 mph would probably have to be moved further towards Lloyd Park.

The Croydon trams have sanding gear, as do all other trams I know about in any detail. I was involved in a detergent on rail test of a Croydon tram when new, and it was very difficult to persuade it to slip even with the sanders isolated.

 

[neonison]

I also believed they had such a strong adhesion that leaf mulch would not affect them significantly.

Given that they use an electromagnetic force to clamp themselves to the rail, instead of relying on gravity as the wheels do: would it be possible that even when adhesion is low and they cannot prevent sliding through sideways friction, that they would still exert enough downward EM force to prevent a derailment? Has there ever been research into that possible effect?

Poor adhesion previously investigated by RAIB:

https://www.gov.uk/raib-reports/collision-between-two-trams-at-shalesmoor

 

[edwin_m]

If the tram is coming up to a blind corner, the line of sight is indeed reduced to a much shorter distance than in other locations.

The tram driver is expected to slow down to the degree of being able to stop if he suddenly sees a shopping trolley on the tracks around the corner (for example) -- essentially the tram is supposed to be driven as though the track that remains unseen just doesn't exist.

This is in contrast to the mainline railway, where a driver is provided with a guarantee (via the signalling systems) that a specific stretch of track ahead is definitely clear, even if it it not yet in sight. This is known as Movement Authority.

All true, and furthermore, the permitted speeds on the speed signs must take account of available sightlines so if the driver is within these speeds it will be possible to stop within the sighting distance using the service brake. Speeds may have to be reduced if visibility is very poor, such as in thick fog.
--- old post above --- --- new post below ---

interesting account of presentation of initial report to H of C

http://www.croydonadvertiser.co.uk/...-explains-mp/story-29894756-detail/story.html

worth a read as it suggests regs might need to be modified where the tram is effectively operating as a train and automatic braking might be needed on the "train "type tracks.

Good find. Unlike most MPs on most subjects, Mr Barwell appears to have good knowledge of trams. He didn't recommend automatic braking, just said it was something for the enquiry to consider.

 

[adrock1976]

Something that has cropped up in my mind here regarding reducing speed on the approach to the aforementioned curve.

Would it be possible to have that section of the overhead wires converted to neutral, so as the tram does not draw power just before exiting the tunnel, so as it would be coasting on the approach to the curve?

Or would this not be a good idea (I'm not an expert on tram technology) if the braking system needs a continuous supply of power from the overhead wires?

 

[Domh245]

Something that has cropped up in my mind here regarding reducing speed on the approach to the aforementioned curve.

Would it be possible to have that section of the overhead wires converted to neutral, so as the tram does not draw power just before exiting the tunnel, so as it would be coasting on the approach to the curve?

Or would this not be a good idea (I'm not an expert on tram technology) if the braking system needs a continuous supply of power from the overhead wires?

The trams can brake regeneratively, but they also have resistor banks on the roof so that they can dump the braking energy if the overheads can't accept any more, and also have disk brakes. But installing a neutral section seems like a lot of faff for no benefit, the driver could still coast through the section at high speed if they wanted.

 

[AlterEgo]

Something that has cropped up in my mind here regarding reducing speed on the approach to the aforementioned curve.

Would it be possible to have that section of the overhead wires converted to neutral, so as the tram does not draw power just before exiting the tunnel, so as it would be coasting on the approach to the curve?

Or would this not be a good idea (I'm not an expert on tram technology) if the braking system needs a continuous supply of power from the overhead wires?

There are a couple of issues there. Firstly, on that tunnel straight the trams are pretty much flat out at 50mph. Going into neutral for a couple of hundred yards isn't going to reduce the speed significantly. Secondly, you'd need a pretty long neutral section to overcome that - a tram, coasting, might take half a mile to stop (I'm not sure of the gradient of that tunnel straight but I assume it to be flat or nearly flat). A long neutral section means a tram can't voluntarily stop in that section or it'll risk being stranded.

I think that key to the investigation are the following issues:

- Why did the driver not brake more than the very small amount recorded? At which point in that magic 180m did he brake? Was it an initial brake application followed by nothing? (This would point towards a medical blackout) Or, was it a desperate, last-minute intervention, pointing towards a fatal loss of concentration, distraction, or microsleep?

- If driver fatigue, inattention or distraction are an issue, was this a one-off or has this happened on several occasions, as possibly alleged? Have other drivers done likewise?

- Crashworthiness: so many questions! - why did seven people die? Initially, I was astonished to learn that even a single person lost their life. It seems that these fatal injuries were caused by people falling out of windows and becoming trapped underneath the tram. Were those people standing or sitting? How did they fall out? Did they die because they were simply crushed by the weight of the tram, or was it because they were scraped along the ballast and rails at an angle? Is the risk of those injuries greater if the tram derails on a segregated RoW with the uneven surface of exposed rails and ballast, rather than on a street section where the road would lie flush with the side of the tram? What about glazing standards? Are the windows too large, exposing passengers to the risk of falling out of them?

 

[Greybeard33]

Good find. Unlike most MPs on most subjects, Mr Barwell appears to have good knowledge of trams. He didn't recommend automatic braking, just said it was something for the enquiry to consider.

Nevertheless, Gavin Barwell MP was not entirely accurate when he said that:

...trams cannot have the same kind of signalling systems as trains.

When Manchester Metrolink first opened in 1991, the ex-heavy rail lines were re-signalled with Track Circuit Block signalling, 2-aspect colour lights and Automatic Tram Stop (ATS) equipment that applies full hazard brake in case of a SPAD, stopping the tram within the signal overlap.

Since then the pendulum has swung the other way, and it seems that the ORR encouraged TfGM to start its programme of migrating the segregated Metrolink lines to line of sight operation like the other second generation tram systems. The block signals and ATS beacons are being progressively removed (with lower speed restrictions imposed in some areas). The human factors justification is apparently that it is undesirable for drivers to have to cope with the switch between line of sight driving on street and Movement Authority driving on segregated sections (although this will be inevitable on tram-train routes).

Road design provides numerous features to mitigate the risk of accidents caused by loss of driver concentration - sweeping curves rather than long fast straights, large advance warning signs, chevrons and speed limits on the approach to sharp bends or other hazards and, in some places, radar signs that flash in case of excessive speed. The need to continually steer and look out for conflicting traffic and roadside hazards also helps road vehicle drivers to maintain concentration.

Perhaps, where ex-heavy rail lines have been converted to line of sight tramways or guided busways, more could have been done to add road-like features that would assist drivers in maintaining concentration and situational awareness?

 

[507 001]

Nevertheless, Gavin Barwell MP was not entirely accurate when he said that When Manchester Metrolink first opened in 1991, the ex-heavy rail lines were re-signalled with Track Circuit Block signalling, 2-aspect colour lights and Automatic Tram Stop (ATS) equipment that applies full service brake in case of a SPAD, stopping the tram within the signal overlap.

Since then the pendulum has swung the other way, and it seems that the ORR encouraged TfGM to start its programme of migrating the segregated Metrolink lines to line of sight operation like the other second generation tram systems. The block signals and ATS beacons are being progressively removed (with lower speed restrictions imposed in some areas). The human factors justification is apparently that it is undesirable for drivers to have to cope with the switch between line of sight driving on street and Movement Authority driving on segregated sections (although this will be inevitable on tram-train routes).

Road design provides numerous features to mitigate the risk of accidents caused by loss of driver concentration - sweeping curves rather than long fast straights, large advance warning signs, chevrons and speed limits on the approach to sharp bends or other hazards and, in some places, radar signs that flash in case of excessive speed. The need to continually steer and look out for conflicting traffic and roadside hazards also helps road vehicle drivers to maintain concentration.

Perhaps, where ex-heavy rail lines have been converted to line of sight tramways or guided busways, more could have been done to add road-like features that would assist drivers in maintaining concentration and situational awareness?

Indeed. Personally, I would rather keep the TCB signalling, and I know a lot of my colleagues feel the same way. I'd certainly rather have TCB than 'more road-like features'.

It is really no hardship to switch between LoS and MA, as you almost completely change your driving style when you come off the street anyway.

Just a slight correction however, SPADing in a TCB area on Metrolink will result in a full hazard brake application, not full service brake.

 

[adrock1976]

The trams can brake regeneratively, but they also have resistor banks on the roof so that they can dump the braking energy if the overheads can't accept any more, and also have disk brakes. But installing a neutral section seems like a lot of faff for no benefit, the driver could still coast through the section at high speed if they wanted.

There are a couple of issues there. Firstly, on that tunnel straight the trams are pretty much flat out at 50mph. Going into neutral for a couple of hundred yards isn't going to reduce the speed significantly. Secondly, you'd need a pretty long neutral section to overcome that - a tram, coasting, might take half a mile to stop (I'm not sure of the gradient of that tunnel straight but I assume it to be flat or nearly flat). A long neutral section means a tram can't voluntarily stop in that section or it'll risk being stranded.

I think that key to the investigation are the following issues:

- Why did the driver not brake more than the very small amount recorded? At which point in that magic 180m did he brake? Was it an initial brake application followed by nothing? (This would point towards a medical blackout) Or, was it a desperate, last-minute intervention, pointing towards a fatal loss of concentration, distraction, or microsleep?

- If driver fatigue, inattention or distraction are an issue, was this a one-off or has this happened on several occasions, as possibly alleged? Have other drivers done likewise?

- Crashworthiness: so many questions! - why did seven people die? Initially, I was astonished to learn that even a single person lost their life. It seems that these fatal injuries were caused by people falling out of windows and becoming trapped underneath the tram. Were those people standing or sitting? How did they fall out? Did they die because they were simply crushed by the weight of the tram, or was it because they were scraped along the ballast and rails at an angle? Is the risk of those injuries greater if the tram derails on a segregated RoW with the uneven surface of exposed rails and ballast, rather than on a street section where the road would lie flush with the side of the tram? What about glazing standards? Are the windows too large, exposing passengers to the risk of falling out of them?

Thanks very much to you both for clearing that up for me.

I'll sit back and wait for the next report from RAIB to come out at some point in the future.

 

[Antman]

From the TfL website:


Before services resumed, and in accordance with advice in the Rail Accident Investigation Branch's (RAIB) interim report into the derailment, additional speed restrictions and associated signage have been introduced near Sandilands.

These precautionary measures have also been implemented at three other locations on the tram network.

All tram drivers were fully briefed prior to resumption and additional staff from TfL and FirstGroup will be available across the network to provide support and information for customers and employees. FirstGroup will also be carrying out enhanced speed monitoring across the tram network.

Engineers have repaired all track and other equipment and have run trams over the repaired section. We have carried out a thorough safety assessment and taken the advice of an independent panel of tram experts as part of a rigorous safety assurance process before resuming the service.

Support is being offered to the families of those who lost their lives or who were injured. This includes the services of our Sarah Hope line, which offers practical, financial and other support.

Mike Brown MVO, London's Transport Commissioner, said: 'Our thoughts are with everyone affected by what happened last Wednesday, and we are doing everything we can to ensure that they receive support at this incredibly difficult time.

'Following the RAIB's interim investigation, we have implemented their advice on additional speed restrictions and signage and carried out a rigorous safety assessment prior to restoring services earlier today.'

Anyone with concerns or comments about the tram service can contact us 24 hours a day on 0343 222 1234.

The Sarah Hope Line is run by dedicated staff to offer practical, financial and other help as needed. The service can be reached on 0343 222 5678, or [email protected].

 

[MikeWh]

And some extra bits from the Press Release:

Tram services had been suspended between East Croydon and Harrington Road, Addington Village and Elmers End to enable the British Transport Police and RAIB to thoroughly investigate the site and to allow TfL engineers to carry out repair work.

As a precautionary measure, and in keeping with the RAIB’s advice, TfL has imposed a further speed restriction before the existing 20 kmph section near Sandilands. A new 60kmph limit has been introduced between the 80kmph limit and the 20kmph limit. The 20kmph limit has also been brought closer to the exit of the tunnel.

Similar restrictions have been implemented at three further locations on the tram network: in the opposite direction between Sandilands and Lloyd Park; on the approach to Sandilands from the Beckenham Junction / Elmers End branch; and on the bend between Birkbeck and Harrington Road.

The independent panel of tram experts was convened by UK Tram, the organisation representing the tram industry in the UK. These experts included an ex RAIB & HMRI (Her Majesty’s Railway Inspectorate) inspector, the ex UK Operations Managing Director for Centro, the ex Managing Director of London Rail, an expert in driver training and the General Manager of UK Tram with an engineering and operations background.

 

[Taunton]

I think the issue of railway-type operation of trams under tram-type regulations will come under scrutiny.

Ironically, old type tramways had this speed on junctions issue cracked a century ago. That was when automatic electric points at road junctions first came along for them. They were operated by a switch connected to the overhead wire, if the tram coasted through the points remained for the straight, if it was taking power they would change for the turn to the side road, typically after having braked first and then moving forward slowly under power. There was the danger of the driver forgetting and passing at speed with the power on, but a further bit of (100 year old) electric switch gizmo-ry was that if the tram collector was in the overhead switch for less than say two seconds, equalling more than a fast walking pace, the points did not change from the straight. This was done by a coil and relay in the circuit. A further switch on the turnout line just beyond the points restored them to the straight route immediately afterwards.

Now the junction in Croydon could be fitted with a straight through set of tracks between the two branches, because an overspeed coming from Beckenham could also happen, and such a traditional device could be fitted. If too fast, the tram would just be sent straight ahead. I'm sure the electrics can be greatly improved 100 years later. It also means that the tram can maintain full speed up to the normal braking position for the junction.

There were several proprietary manufacturers of the time. Here's a (5mb download) brochure of the period about one. It handled all sorts of failure issues, short circuit up in the switch, or tram stopped under the switch.

http://www.cotma.net.au/library/signals/Collins Automatic Point Turner Sheet 18.pdf

 

[gavin]

"Shocking video shows tram driver SLEEPING at the wheel while doing 40mph on same line as horror Croydon crash where seven passengers were killed"

https://www.thesun.co.uk/news/22163...don-crash-where-seven-passengers-were-killed/

 

[AlterEgo]

"Shocking video shows tram driver SLEEPING at the wheel while doing 40mph on same line as horror Croydon crash where seven passengers were killed"

https://www.thesun.co.uk/news/22163...don-crash-where-seven-passengers-were-killed/

Wow. That is incredible.

 

[Robertj21a]

Oh dear, that looks like it's going to prove to be genuine.