The extent to which slippery rails can cause trains to be unable to stop?

[king_walnut]

Moderator note: Split from https://www.railforums.co.uk/threads/incident-at-enfield-town-12-10-21.223368/

I would hope so, my gut feeling is railhead contamination of some kind but we will see.

Absolutely not.

That train is several coaches long. The rails and the wheels would have to be entirely made of black ice for that train to be unable to stop.

Trains were pulling in and out of that platform all morning with no problems.

 

[507 001]

Absolutely not.

That train is several coaches long. The rails and the wheels would have to be entirely made of black ice for that train to be unable to stop.

Trains were pulling in and out of that platform all morning with no problems.

Erm, no, it doesn’t. One coach sliding would be a significant reduction in brake force.

 

[king_walnut]

Erm, no, it doesn’t. One coach sliding would be a significant reduction in brake force.

A train pulled in to Enfield Town Platform 2 at 07:54 with no problems, it then pulled out of Platform 2 at 08:08 with no problems.

Then just 6 minutes later, according to you, the rails became so slippy that an 8 coach train slid so hard into the buffer stops that the entire front coach is in the air.

In an area with no vegetation surrounding it. After the railhead cleaning train was there at 3am. With an ambient temperature of 10 degrees celcius.

Amazing.

 

[GB]

A train pulled in to Enfield Town Platform 2 at 07:54 with no problems, it then pulled out of Platform 2 at 08:08 with no problems.

Then just 6 minutes later, according to you, the rails became so slippy that an 8 coach train slid so hard into the buffer stops that the entire front coach is in the air.

In an area with no vegetation surrounding it. After the railhead cleaning train was there at 3am. With an ambient temperature of 10 degrees celcius.

Amazing.

Not saying railhead conditions were the cause but vegetation is not the only cause of poor railhead conditions.

 

[king_walnut]

Not saying railhead conditions were the cause but vegetation is not the only cause of poor railhead conditions.

What causes railhead conditions to change from absolutely fine to so bad that a train smashes into the buffer stops in the space of 6 minutes? Maybe a team of people walking along the four foot pouring engine oil all over it?

Or is it more likely to be literally anything else?

 

[GB]

No need to be facetious. Merely pointing out that when talking about poor railhead conditions, vegetation is not the only source.

As I said, I am not saying railhead conditions were or were not the cause.

 

[507 001]

A train pulled in to Enfield Town Platform 2 at 07:54 with no problems, it then pulled out of Platform 2 at 08:08 with no problems.

Then just 6 minutes later, according to you, the rails became so slippy that an 8 coach train slid so hard into the buffer stops that the entire front coach is in the air.

In an area with no vegetation surrounding it. After the railhead cleaning train was there at 3am. With an ambient temperature of 10 degrees celcius.

Amazing.

There’s this weird phenomenon called rain. That could easily make the rails that slippery within 6 minutes. In fact the less it rains the more slippery it gets.

It doesn’t take every single wheel set to be sliding to result in an incident like this.

A friend of mine once collided with another M5000 at St Peter’s Sq in Manchester when trying to stop behind it on the platform. SPS is uphill in the direction concerned, and the first M5000 didn’t have any issue stopping. Guess what the OTMR showed on the second M5000? But you carry on trying to blame the driver, which is the only logical explanation for your demeanour. At the end of the day, unless you were in the cab with the driver then you cannot rule anything out completely.

I’d also suspect that the collision here was probably around 10mph…

 

[Darandio]

A train pulled in to Enfield Town Platform 2 at 07:54 with no problems, it then pulled out of Platform 2 at 08:08 with no problems.

Then just 6 minutes later, according to you, the rails became so slippy that an 8 coach train slid so hard into the buffer stops that the entire front coach is in the air.

In an area with no vegetation surrounding it. After the railhead cleaning train was there at 3am. With an ambient temperature of 10 degrees celcius.

Amazing.

The train 6 minutes earlier could have contaminated the railhead. It's happened before and been a contributing factor in an accident, in that case the contamination had been carried from a location two miles away. No vegetation in the immediate area doesn't mean no contamination.

 

[theking]

Erm, no, it doesn’t. One coach sliding would be a significant reduction in brake force.

Sorry but that's not adhesion related it's either a train fault or the most likely driver error.

These are not units from the 60's they've got one of the most advanced wsp systems on the network.

 

[507 001]

Sorry but that's not adhesion related it's either a train fault or the most likely driver error.

These are not units from the 60's they've got one of the most advanced wsp systems on the network.

I’m not trying to say 100% either way, I was merely trying to respond to the assertion that it would require all wheels to be locked to result in an overrun to that extent, which isn’t necessarily true.

I’m also well aware of the nature of these units and that they’re very much not 60s units, but no WSP system is infallible (we’re getting close to unsinkable titanic territory here) and I highly doubt any WSP system will be able to maintain the requested brake force whilst intervening in a slide, no matter how advanced it is.

You may well be correct, it may well be the driver’s fault, then again it may not! Remember the voyager that tried to escape across the concourse t Chester a few years ago? So let’s keep an open mind and try not to have him or her hung drawn and quartered before tea time.

 

[physics34]

What causes railhead conditions to change from absolutely fine to so bad that a train smashes into the buffer stops in the space of 6 minutes? Maybe a team of people walking along the four foot pouring engine oil all over it?

Or is it more likely to be literally anything else?

Very fine drizzle

 

[Murph]

Absolutely not.

That train is several coaches long. The rails and the wheels would have to be entirely made of black ice for that train to be unable to stop.

Trains were pulling in and out of that platform all morning with no problems.

It has happened before. Class 221s are "several coaches long", but one hit the buffers in Chester due to poor adhesion (and a design / engineering issue with the sanders on 22x units):

Report 26/2014: Buffer stop collision at Chester station

Buffer stop collision at Chester station, 20 November 2013.

www.gov.uk

A common factor between the 710 in this incident and the class 221 would be disc brakes. Normally that gives superior braking performance, but has the downside that it doesn't clean contamination from the wheels.

 

[millemille]

Sorry but that's not adhesion related it's either a train fault or the most likely driver error.

These are not units from the 60's they've got one of the most advanced wsp systems on the network.

While not speculating on the cause of the accident I feel I should point out that WSP has little or nothing to do with a trains ability to stop in low adhesion conditions, once the brake demand exceeds the coefficient of friction between wheel and rail the train will only stop at a rate that the coefficient of friction* allows. WSP's purpose is to prevent the creep ratio (the difference between wheel rotational velocity and vehicle speed over ground) from exceeding a predetermined level at which tread damage occurs and allowing the brakes to remain applied.

I've had this very conversation, as I design sander systems for a living, with many people in the industry who should know better over the last few years. The steel of the wheel and rail don't know or care what WSP and brake control systems are trying to stop the train.


*typical clean dry wheel and rail has a coefficient of friction of 0.6, which allows braking of up to 60% of g. Normal timetabling relies on achieving brake step 2 which is 6% of g, requiring a coefficient of friction of 0.06. Railhead coefficient of friction <0.01 have been measured on UK infrastructure, meaning wheel slide occurs as soon as any brake demand is made which exceeds 1% of g which makes stopping difficult when brake step 1 is typically 3% of g.

Well, if we're having a new thread...

Report 18/2011: Station overrun incident at Stonegate

Station overrun incident at Stonegate, East Sussex, 8 November 2010.

www.gov.uk

Report 26/2014: Buffer stop collision at Chester station

Buffer stop collision at Chester station, 20 November 2013.

www.gov.uk

 

[gazr]

Has it already been mentioned that a Network Rail train (loco hauled) visited the platform (2) around 3AM the morning of the incident? Sat in the platform for 10 minutes. Coincidental? https://www.realtimetrains.co.uk/service/gb-nr:H13806/2021-10-11/detailed

 

[Horizon22]

If it’s Derby it’s probably the national measurement train. RHTT normally run from more local depots, but there’s every chance it could be that.

 

[GB]

Even if the RHTT made a visit its doubtful it would have sprayed a dead end station as you can’t spray below 20mph.

 

[scrapy]

Even if the RHTT made a visit its doubtful it would have sprayed a dead end station as you can’t spray below 20mph.

Yes whenever I've seen them go into a terminal station they simply go behind the starting signal to turn round rather than the full platform length, and don't spray the platform area.

 

[D6130]

If it’s Derby it’s probably the national measurement train. RHTT normally run from more local depots, but there’s every chance it could be that.

Hmmm......if that were the case, there is the potential for an oil leakage onto the rail from one of its (class 37 or 43?) locomotives.

 

[Horizon22]

Hmmm......if that were the case, there is the potential for an oil leakage onto the rail from one of its (class 37 or 43?) locomotives.

Highly unlikely.

 

[Llama]

I'm surprised to hear that the coefficient of friction can be as high as 60%g - surely that's an absolute maximum for a 'lab conditions' clean, dry, freshly-lathed wheel on rail? From adhesion related incident reports I've read in the past "at least 20%" was given as a more likely figure for a dry wheel & rail without contamination, and that generally being why on most modern rolling stock enhanced emergency brakes are nominally 12%g, service braking generally being up to 9%g. Obviously I'd expect that CoF figure to be conservative with a fairly wide margin but it's surprising to hear it'd be up to three times that value.

On some of our new units it was found pretty soon after introduction (early 2019 leaf-fall season) that the configuration of the sander/WSP system was totally wrong and that they had been set up with a far looser 'wheel slow-up' tolerance before the auto-sanders worked than was realistic which led directly to adhesion related incidents. I forget the figures now but the reduction of wheelset rotational speed vs train speed before sand was deployed (sanding under service braking only works by the auto-sanders, not manually by the driver) was something like 60-70% when it should've been something like 12%. The software on the units were modified pretty quickly to reduce the figure to what it should've been.

I posted a link somewhere else although the original thread has been split - the three RAIB reports shown under 'autumn review' - Esher, Lewes and the adhesion review report - should be required reading for drivers and others on the subject of low adhesion in my opinion. Do a search for 'RAIB adhesion' and they'll appear.

 

[seagull]

I've personally experienced a FOUR MILE slide of a train, formed of "several coaches" with working sanders and WSP, albeit aggravated by the fact the train went into full emergency braking, at 70mph, downhill and in the autumn.
Neither rails nor wheels were manufactured from black ice, that is one thing I do know.

 

[O L Leigh]

I do think it's something that people from outside of the industry fail to understand. My parallel is driving on ice with normal summer tyres; no grip for power, braking or steering. Leaf contamination, contamination from road vehicles at crossings (particularly after gritting), mist/fog/fine rain, fine dust on the railhead can all affect adhesion.

Things have got better since railhead treatment trains (RHTT) were introduced, but it still the case that you can get very poor adhesion which can suddenly rob you of your ability to stop. I've had a few "brown trouser" moments because of this, and not all of them during the autumn, so I always exercise caution.

The Enfield Town branch always used to be very bad for low adhesion. In the bad old days you'd creep around Bury Street Jn at 15mph (linespeed is 50mph) and apply the brake straight away. The unit would go into an all-wheel slide immediately and accelerate down the gradient. You'd wave at all the punters on the platform at Bush Hill Park as you silently glide through and then finally come to a stand halfway up the hill towards Enfield Town.

 

[edwin_m]

I'm surprised to hear that the coefficient of friction can be as high as 60%g - surely that's an absolute maximum for a 'lab conditions' clean, dry, freshly-lathed wheel on rail? From adhesion related incident reports I've read in the past "at least 20%" was given as a more likely figure for a dry wheel & rail without contamination, and that generally being why on most modern rolling stock enhanced emergency brakes are nominally 12%g, service braking generally being up to 9%g.

If it can go that high it's only really of academic interest. To exploit that level of adhesion, braking systems would have to be much more powerful and passengers would have to be strapped in. Even then it could only be used if that level of adhesion was actually available, which would be rare if ever. "At least 20%" allows for the normal braking rates mentioned and provides a bit of a margin.

In case anyone doesn't know, the coefficient of friction must be at least as much as the demanded brake application, expressed as a percentage of gravity.

 

[millemille]

I do think it's something that people from outside of the industry fail to understand. My parallel is driving on ice with normal summer tyres; no grip for power, braking or steering. Leaf contamination, contamination from road vehicles at crossings (particularly after gritting), mist/fog/fine rain, fine dust on the railhead can all affect adhesion.

Gatwick airport is a known low adhesion area all year round whenever the railhead gets wet, attributed to the unburnt fuel in the jet exhaust from the planes flying over the tracks at low level.

If it can go that high it's only really of academic interest.

Indeed, it is the benchmark figure only and only seen on test rigs.

I've personally experienced a FOUR MILE slide of a train, formed of "several coaches" with working sanders and WSP, albeit aggravated by the fact the train went into full emergency braking, at 70mph, downhill and in the autumn.
Neither rails nor wheels were manufactured from black ice, that is one thing I do know.

My wife, when she as driving for Southeastern donkeys' years ago, had multi mile slide when she put the brake in step 1 on the way down to Dover in a 365 on a training trip. She said the speedo wound down to zero but she could feel the train accelerating, it all went very quiet in the cab and the DI dropped the 2nd man window and they could hear the dump valves chattering away. Nothing to do but sit there and hope it came back to her while the DI called the bobby and told them they were sliding.

 

[GC class B1]

The straightforward comparison between braking rate as a percentage of g and the required wheel/rail coefficient of friction is too simplistic. It assumes that all the wheel loads are equal. Wheel loads (weight) will vary according to a lot of factors and therefore some wheelsets will be more likely to slide than others with the same wheel/rail coefficient of friction.

 

[ComUtoR]

We have had slip throughs where the unit goes out the platform faster than it came in. We had one where the Driver came in around 12mph and went out the other side 20 odd mph.

A train hitting a buffer even at 5mph makes significant damage.

 

[edwin_m]

The straightforward comparison between braking rate as a percentage of g and the required wheel/rail coefficient of friction is too simplistic. It assumes that all the wheel loads are equal. Wheel loads (weight) will vary according to a lot of factors and therefore some wheelsets will be more likely to slide than others with the same wheel/rail coefficient of friction.

Frictional force is coefficient of friction multiplied by normal reaction. Normal reaction in this case is essentially the weight on the wheels in question.

Acceleration is force divided by mass. Put these together and the mass cancels out, the deceleration (as % of gravity) is the same as the coefficient of friction (between wheel and rail) needed to achieve it.

There is also friction between the brake pads and discs. The decelerating force achievable depends on the normal reaction applied by the pads on the discs, and on the coefficient of friction here, which will be greater than between steel and steel because the pads are designed to maximise it. However most modern friction brake systems will adjust the pad force to give a reasonably constant deceleration independent of the weight of the vehicle, including sensing the weight of the passengers, so that a lighter vehicle won't slide before the rest of the train, and the same deceleration is achieved regardless of load. Electric brake systems apply a retarding force using the motors, but will adjust it in the same way.

 

[GC class B1]

Frictional force is coefficient of friction multiplied by normal reaction. Normal reaction in this case is essentially the weight on the wheels in question.

Acceleration is force divided by mass. Put these together and the mass cancels out, the deceleration (as % of gravity) is the same as the coefficient of friction (between wheel and rail) needed to achieve it.

There is also friction between the brake pads and discs. The decelerating force achievable depends on the normal reaction applied by the pads on the discs, and on the coefficient of friction here, which will be greater than between steel and steel because the pads are designed to maximise it. However most modern friction brake systems will adjust the pad force to give a reasonably constant deceleration independent of the weight of the vehicle, including sensing the weight of the passengers, so that a lighter vehicle won't slide before the rest of the train, and the same deceleration is achieved regardless of load. Electric brake systems apply a retarding force using the motors, but will adjust it in the same way.

The point I am making is that the weight of the vehicle only cancels out if the weight is distributed evenly. It is correct that the brake force generated will take into account the vehicle weight using the air suspension pressure which is generally averaged across the two bogies. However the individual wheelset weights will vary according to the vehicle weight distribution, features of the suspension and weight transfer under braking. There will also be some difference in the performance of individual brake actuators, callipers and pads.
This explains why only some wheelsets will lock when others on the same vehicle don’t.

 

[millemille]

@GC class B1 you've edited your post as I was typing my reply.

RSSB (presumably you mean GMRT2461?) don't "mandate" anything with regards to the location of sanders on a vehicle. They make recommendations.

There are several reasons why it is preferable to not sand underneath the leading axles, low voltage track circuit assured activation being one but the more important one is you can't sand underneath the leading wheelsets with automatic sanders otherwise you end up in a horrendous feedback loop; brakes apply, WSP activity occurs, sander activates, sand is laid under leading wheels, WSP activity stops, sander stops, WSP activity starts, sander activates, sand is laid under leading wheels, WPS activity stops, sander stops, WSP activity occurs.....and so on and so on. You have to have at least one axle that is running on un-sanded track in order to get a WSP activity signal that is representative of the un-sanded track condition.

 

[43066]

While not speculating on the cause of the accident I feel I should point out that WSP has little or nothing to do with a trains ability to stop in low adhesion conditions, once the brake demand exceeds the coefficient of friction between wheel and rail the train will only stop at a rate that the coefficient of friction* allows. WSP's purpose is to prevent the creep ratio (the difference between wheel rotational velocity and vehicle speed over ground) from exceeding a predetermined level at which tread damage occurs and allowing the brakes to remain applied.

I've had this very conversation, as I design sander systems for a living, with many people in the industry who should know better over the last few years. The steel of the wheel and rail don't know or care what WSP and brake control systems are trying to stop the train.


*typical clean dry wheel and rail has a coefficient of friction of 0.6, which allows braking of up to 60% of g. Normal timetabling relies on achieving brake step 2 which is 6% of g, requiring a coefficient of friction of 0.06. Railhead coefficient of friction <0.01 have been measured on UK infrastructure, meaning wheel slide occurs as soon as any brake demand is made which exceeds 1% of g which makes stopping difficult when brake step 1 is typically 3% of g.

Well, if we're having a new thread...

Report 18/2011: Station overrun incident at Stonegate

Station overrun incident at Stonegate, East Sussex, 8 November 2010.

www.gov.uk

Report 26/2014: Buffer stop collision at Chester station

Buffer stop collision at Chester station, 20 November 2013.

www.gov.uk

Excellent post, thanks.

The Stonegate slide has to be one of the longest ever recorded and when people sneer about “leaves on the line” is a great way of demonstrating how much of an issue adhesion (or lack thereof) can be. Especially when combined with unsatisfactory maintenance.