Detecting leaves on line and other hazards

[KeithMacdonald]

According to Engineering & Technology:

 

Quote

A system that allows trains to detect “low adhesion hazards”, such as leaves on the line or other issues that could cause the rail equivalent of black ice, has been developed by researchers from Loughborough University.

 

Sounds good, but why?
 

Quote

 

Low adhesion is caused by the contamination of railways lines by biological, chemical and physical factors, some of which cannot be easily monitored or controlled. The estimated overall cost of low adhesion to the UK railway industry is estimated at £350m each year, according to the Rail Safety and Standards Board. A minimum level of adhesion is essential for reliable braking and traction performance, especially for maintaining safety and limiting delays. Changes in adhesion can be very localised, unpredictable and transient. Poor adhesion experienced by one train may not affect following trains at the same location. The newly developed system will detect low adhesion hot spots in real-time and create an up-to-date map of the UK’s network which shows where any hazards might be. The hope is that the information will allow network operators to react quickly to potential risks, allowing services to run more safely and smoothly.

 

 

https://eandt.theiet.org/content/articles/2022/01/trains-given-ability-to-detect-leaves-on-the-line-and-other-hazards/?utm_source=Adestra&utm_campaign=New EandT News - Automation FINAL - NON MEMBER&utm_medium=Newsletters - E%26T News&utm_content=E%26T News - Non-Members&utm_term=3550444

 

How times and the technology have changed for track (and road) clearance.

Just twenty or so years ago, for example, early-morning BBC Radio Two would regularly report on the A39 Cockbridge to Tomintoul road, where the legendary silver-haired postmistress Mrs MacKay was allegedly (and single-handedly) clearing the snow with her silver-handled shovel. Sometimes, they said, she would be out shovelling snow as early as June. Now, of course, doing that would be a "health and safety" issue with risk assessments and due dilgence appraisals.

 

So anyway of automating the clearing of track (and roads) is very welcome.

 

How did they used to do it on tracks?

[BR traction instructor]

The leaf fall season would commence October ish and last until around Christmas in the Peak District. An old DMU trailer with Sandite (gritty paste) gear on board (large tank with discharge pipes dropping down to just above the rail surface and a pump) would be hauled or propelled by a loco (typically a 25 in the 70s/80s) as required and usually overnight, discharging Sandite paste onto the rail in known problem areas. Less of a problem when lineside tree/vegetation clearance was a responsibility taken seriously by British Rail. 
 

Low speeds were used when applying Sandite. Shed staff from Buxton crewed their Sandite unit when operating.

 

BeRTIe

Edited January 18, 2022 by BR traction instructor

[Wheatley]

Is the question how did (do) they detect and manage poor rail conditions ?

 

Three main methods:

 

1.Blanket instruction to drivers to brake earlier and lighter, with additional instructions at known hotspots (known from experience). Causes delay. 

 

2. Driver reports 'unusually severe conditions' for a particular  location to the signaller, who must then caution trains until a notice can be wired/faxed/emailed out to Signing On Points. Causes more delay. 'Normally severe conditions' for a particular location are covered by (1) above. 

 

3. Driver has an adhesion related incident - a SPAD or station overrun - and gives poor adhesion as the cause. Control may require the following service to make a 'controlled stop' and report back - if that slides past as well then driver 1 is usually exonerated and driver 2 is not admonished either. Network Rail staff have some hand held widgets now which measure railhead conditions but I've no idea how they work. In my day it was done by looking at it and deciding whether it looked more black than silver. Causes a lot of delay. 

 

If the train can detect it as it goes along then that's potentially a game changer, especially as the technology already exists for trains to talk to each other and share that data.  

[Edwin_m]

There was a Low Adhesion Warning System under development in the former BR Research around 20 years ago.  

[Oldddudders]

Leaves on the line is a good music-hall excuse for the media. And it is a relatively modern phenomenon. 

 

In times past, when people were a cheap commodity, Permanent Way gangs roamed their patch daily, inspecting the track, knocking in the keys on bullhead and keeping lineside vegetation to a minimum, a process often helped by passing steam locos. But staff costs grew, and long-welded rail came along, eliminating the need for such a rigorous inspection regime, so the PW gangs dwindled. And the lineside veg grew and grew, to the general delight of the railway's neighbours, who found it deadened noise and encouraged wildlife. Saplings became trees, and trees shed leaves. 

 

On 1.4.94, when Railtrack separated from the BR Board, the principal contract for maintenance, then still undertaken by BR staff, was RT1. Given the amount of effort that had been put into it, it was surprising that it contained no clause about vegetation clearance or control. 

 

In some respects, Southern Region was ahead of the game, as it already employed oil-spray trains to coat the conductor rail with antifreeze mixture when temperatures were forecast to drop. These proved reasonably easy to convert to use Sandite in the Autumn, although it can occur that the leaves are still in gay profusion and frosts have already started. This keeps depot staff busy doing conversions.

[SM42]

15 hours ago, Wheatley said:

Is the question how did (do) they detect and manage poor rail conditions ?

 

Three main methods:

 

1.Blanket instruction to drivers to brake earlier and lighter, with additional instructions at known hotspots (known from experience). Causes delay. 

 

2. Driver reports 'unusually severe conditions' for a particular  location to the signaller, who must then caution trains until a notice can be wired/faxed/emailed out to Signing On Points. Causes more delay. 'Normally severe conditions' for a particular location are covered by (1) above. 

 

3. Driver has an adhesion related incident - a SPAD or station overrun - and gives poor adhesion as the cause. Control may require the following service to make a 'controlled stop' and report back - if that slides past as well then driver 1 is usually exonerated and driver 2 is not admonished either. Network Rail staff have some hand held widgets now which measure railhead conditions but I've no idea how they work. In my day it was done by looking at it and deciding whether it looked more black than silver. Causes a lot of delay. 

 

If the train can detect it as it goes along then that's potentially a game changer, especially as the technology already exists for trains to talk to each other and share that data.  

 

The hand held widget is an eddy current device, which through magic ( or perhaps physics) can measure the amount of leaf contamination on the rail head. 

The result is fed back to control to allow treatments to be targeted more effectively. 

 

Back in the 90s there was a system installed on some areas where there was a recognised high risk of leaf fall related  adhesion issues. 

 

If a report had been received the signaller could switch it on. 

The result was a large dot matrix style trackside sign was illuminated to warn drivers and this removed the need to stop the  train ( with the SPAD risk that may introduce) and tell the driver. 

 

Andy

[Ian Smeeton]

Big Jim had a piece on his out & about with.... thread about leaves on the line.  Worth a look.

 

I'll try and see if I can find it and add a link.

 

Regards

 

Ian

[Ian Smeeton]

Found it!

 

 

Regards

 

Ian

[The Stationmaster]

I always found it useful to correlate any increase in 'leaveson the line' delays with the date when steam finished on any route and lineside vegetation control was cut back reduced.  thus on. the Wr in the late 1960s 'leaves on the line delays, even in teh West of England, were unheard of but by the 1980s, with 20 years of growth on trees within the railway boundary they were occurring, and frequently in some locations.

 

Because controlled burning had normally been used (except where open cable routes existed) on the WR to control linesode vegetation trees didn't stand much chance of growing so they weren't there to drop leaves.  Elsewhere with electric trains replacing many of teh steam hauled services fromearlier years 'leaves on the line' seemed to predate WR experience with the problem.

 

So there is really a very simple answer to the problem - cut down the trees and restore railway embankments and cutting sides to the way they used to be which was suitable for a wide range of small mammals and various flora etc - all of which were driven away by tree growth.  Most of the people who complain about trees being felled at the railway lineside simply can't (or won't) remember what preceded the green jungle of trees and accompanying undergrowth which has had such an adverse impact on the flora and fauna which once inhabited these areas.

Edited January 19, 2022 by The Stationmaster

[virginhst539]

 

I can only comment on what I have seen but anyway here we go. 

 

So there are your rail head treatment trains, in my area its MPV on one circuit and TnT class 67s on the other. Normally doing a day time run and a night run, 7 days a week. On top of this there are also a dedicated team that covers known trouble spots that will inspect railhead condition, remove leaf fall from the track and scrub and treat contaminated rail head areas. Obviously the ones who will detect contamination issues are the drivers who phone it in and the teams will respond to clear it. 

 

6 hours ago, The Stationmaster said:

 

So there is really a very simple answer to the problem - cut down the trees and restore railway embankments and cutting sides to the way they used to be which was suitable for a wide range of small mammals and various flora etc - all of which were driven away by tree growth.  Most of the people who complain about trees being felled at the railway lineside simply can't (or won't) remember what preceded the green jungle of trees and accompanying undergrowth which has had such an adverse impact on the flora and fauna which once inhabited these areas.

As for the trees its a bit of a catch 21 situation. People want the trains to run with out delay but they also don't want you to cut the trees down near there property which blocks there view of the railway. There are other things to consider such as sites of scientific interest, areas with protected species sightings, nesting birds and tree preservation orders which greatly limits what can be done in these areas.

[AMJ]

As most track has circuitry for the signals that creates a power connection when the steel wheels are on the rails if the train fails to detect the signaling system it should be set to slow down and "report" the location.  I have heard of trains vanishing on some lines as they were not detected.

[The Stationmaster]

20 hours ago, virginhst539 said:

 

 

As for the trees its a bit of a catch 21 situation. People want the trains to run with out delay but they also don't want you to cut the trees down near there property which blocks there view of the railway. There are other things to consider such as sites of scientific interest, areas with protected species sightings, nesting birds and tree preservation orders which greatly limits what can be done in these areas.

That is exactly part of the problem regarding the trees - many people seem to think they've been there forever and are an important part of the natural world.  But in reality the vast majority of them are fairly recent and have totally changed and destroyed the previously long established lineside habitat for flora and fauna, to their detriment.  All you need to do is compare photos of the lineside nowadays with photos of the lineside taken at almost any time during the steam operated life of almost every stretch of railway you choose to look at.

 

One modeller on here was I think a little surprised when I mentioned that his lineside scenery didn't show what actually happened in the period he is modelling - when the land immediately outside the railway fence (and for several decades after) was meadow grazed by cattle.

 

For another example look at the picture below which shows the former SR mainline diving  under the WR main line at St Budeaux am nd loom art the mass of  surrounding both lines.  Now find Stirling road at the top of the photo - when i stayed there with relatives in the late 1950s from a first floor window  I could read the numbers of engines on the SR route provided they were moderately clean and see the numberplates on Western engines although they were too far away (and usually too dirty) to read.  Now even if lots of buildings didn't now obscure the view,  you couldn't even see the railway from there because of the tree growth.  A stark but simple illustration of massive change in only a few decades.

 

[SM42]

11 hours ago, AMJ said:

As most track has circuitry for the signals that creates a power connection when the steel wheels are on the rails if the train fails to detect the signaling system it should be set to slow down and "report" the location.  I have heard of trains vanishing on some lines as they were not detected.

 

This is the biggest risk that the public don't see or hear about ( but certainly see the effects) and it's no music hall joke. One of the main reasons so much effort goes into every Autumn. 

 

Disappearing off the signalling system has implications for  safety  primarily but also for performance.

 

The network loses a signal section, possibly more,   in the safety mitigations following a wrong side track circuit failure.

The disappearing train also gets stopped so that the wheels  can be checked for leaf contamination 

 

The timetable suffers as a result and the S&T spend a not insignificant  amount of time investigating ( in all weathers) to establish  the cause taking them away from their other work. 

 

 

Whilst all this is happening  spare a thought for

 

1. The signaller ensuring  the safety of trains through the affected area, the safety of the team out on the track and all the other trains that they have to keep moving elsewhere on their bit of railway.

 

Throw in a second problem on the patch ( not uncommon) and it can be a thankless task some days.

 

2. The control teams juggling their resources and probably dealing with  several other problems at the same time.

Another thankless task

 

3. The train crew supervisor dealing with crews not being where they should be at the given time as they are late inbound

 

4. The public facing staff getting the flak.

 

5. The passengers who arent getting there when they expect to.

 

 

In some cases the train disappears for only a second or two  which can be easily missed on a busy panel. Fortunately modern signalling  installations have very effective electronic monitoring which can pick up failures that the human eye may not see. 

 

Andy

 

 

[PaulRhB]

On 20/01/2022 at 14:34, AMJ said:

As most track has circuitry for the signals that creates a power connection when the steel wheels are on the rails if the train fails to detect the signaling system it should be set to slow down and "report" the location.  I have heard of trains vanishing on some lines as they were not detected.

Ok there is intelligent infrastructure monitoring on many tracks already and we get calls from control to block work a certain section as a result, making sure the train has passed clear manually before setting routes behind it. Then a MPV or RHTT will treat it as it’s usually too long for manual treatment although in certain cases a manually pushed ‘scrubber pram’ has been used in small areas. 
The system can’t automatically slow down the system as it isn’t linked to the Signalling, it may become possible in the future but not with current security risk levels on the net. An alarm going off and responded to by Control and Signaller works well so far. 
Adhesion levels can fall dramatically without the train disappearing from a track at all too. 
A problem with the idea is the increasing use of axle counters which do not detect the train through the rails at all so they cannot detect contamination at all. 

As many will know we had the collision at Salisbury in October last year so I would keep an eye on the RAIB reports to see what that has to say as it will be a factor considered. 

Edited February 3, 2022 by PaulRhB

[Northmoor]

On 19/01/2022 at 10:57, Oldddudders said:

Leaves on the line is a good music-hall excuse for the media. And it is a relatively modern phenomenon. 

 

In times past, when people were a cheap commodity, Permanent Way gangs roamed their patch daily, inspecting the track, knocking in the keys on bullhead and keeping lineside vegetation to a minimum, a process often helped by passing steam locos. But staff costs grew, and long-welded rail came along, eliminating the need for such a rigorous inspection regime, so the PW gangs dwindled. And the lineside veg grew and grew, to the general delight of the railway's neighbours, who found it deadened noise and encouraged wildlife. Saplings became trees, and trees shed leaves. 

 

On 1.4.94, when Railtrack separated from the BR Board, the principal contract for maintenance, then still undertaken by BR staff, was RT1. Given the amount of effort that had been put into it, it was surprising that it contained no clause about vegetation clearance or control. 

 

In some respects, Southern Region was ahead of the game, as it already employed oil-spray trains to coat the conductor rail with antifreeze mixture when temperatures were forecast to drop. These proved reasonably easy to convert to use Sandite in the Autumn, although it can occur that the leaves are still in gay profusion and frosts have already started. This keeps depot staff busy doing conversions.

This is all true, but there is an additional factor which is the performance of modern units.  There may have always been low adhesion risks in steam days, but they simply could not put down sufficient accelerative or braking force to exceed the available adhesion.  If there are times of the year when available adhesion drops (to lower levels than historically) and deceleration rates are higher than previously, at some point - increasingly often - the values will equate and a slide will occur.

And I agree it might be a media joke but The Autumn Issue isn't really about leaves on the line, it's about entire trees on the line.  Even with a couple of hundred tons of train, you really don't want to hit one.

[PaulRhB]

22 minutes ago, Northmoor said:

There may have always been low adhesion risks in steam days, but they simply could not put down sufficient accelerative or braking force to exceed the available adhesion. 

Unfortunately not always true  I’ve driven a steam loco, albeit narrow gauge, where we quite easily achieved a speed where the wheels locked under braking due to damp track in the autumn. I was working the handbrake to keep the wheels slowly rotating to prevent flats but all I was achieving was checking the speed of the train and couldn’t stop it on a downhill gradient. I wasn’t worried about signals as I was nowhere near one but I was concerned about the curve at the bottom of the grade. Remember you can get good rail conditions in some places and poor in others so accelerating and braking can suddenly change especially on the mainline due to distances covered. There are known poor places but if previous stops haven’t been bad you don’t expect a total change if the conditions haven’t changed. If it rains it alerts you to potential changes but lack of sun at the right time can leave moisture on the rails too. 
With modern units there is technology looking at wheelslip but all wheels might stop in a slide and other factors such as train length and type also affect braking. 

[Edwin_m]

Traditionally wheelslide protection systems worked by comparing the speed of the wheels, and this is probably still quite effective because the front wheels tend to clean the rails to some extend for later ones, and also if the train suddenly hits a bad patch (such as oil spilled on the rail) the front will react before the rear reaches it.  Modern sanding systems also apply to the second bogie so the front one can be used to detect poor adhesion.  But I assume modern WSP systems would detect and react to a sudden non-feasible drop in speed on all the wheels they are comparing.  

[PaulRhB]

59 minutes ago, Edwin_m said:

Traditionally wheelslide protection systems worked by comparing the speed of the wheels, and this is probably still quite effective because the front wheels tend to clean the rails to some extend for later ones, and also if the train suddenly hits a bad patch (such as oil spilled on the rail) the front will react before the rear reaches it.  Modern sanding systems also apply to the second bogie so the front one can be used to detect poor adhesion.  But I assume modern WSP systems would detect and react to a sudden non-feasible drop in speed on all the wheels they are comparing.  

Possibly on some but certainly not widespread on older stock including the 158/9 fleets. As I said above the length of train can make a big difference and it’s not unknown to get an all wheels slide on a 158. 

[Supaned]

All the technology in the world won't alter the basic physics that the point of contact between a wheel and a rail is very small and once there is insufficient friction between the two to grip , a slide will occur.

 

Granted this is exacerbated by newer trains which are lighter in weight , disc braked and in some cases also fitted with regenerative or dynamic braking which is great in good rail conditions but more of a hinderance in low adhesion - I think my record for a slide so far is on the Up line from Worting Junction to Basingstoke - brake applied at Parlour Gate LC (4 signal sections earlier than the "normal" braking area, all wheels locked and slid, brake in emergency stopping not far off the signal protecting Basingstoke station. I then crept into the station at approx 10 mph and nearly slid out the other end....definately NOT a laughing matter.

 

I'm not quite sure exactly what some technology that "detects" rail contamination is meant to do? if it reduces braking force on a train to compensate than you have additional risk factors associated with that when a driver applies the brakes and they don't work as expected , yet it could give some sort of speed warning which is then a performance issue, and in truth trains can slide at any speed if the conditions are that bad, so that's not really a solution either.

 

 

[PaulRhB]

I remember two of my colleagues were involved in a huge timetable rewrite back in 2005 to increase reliability. Mainly it increased dwell times at stations and avoided close timings at conflict points. Over the last 17 years I’ve seen all those performance recovery buffers cut and increasingly 1 min station stops that are difficult enough to unload, load and dispatch let alone allow resilience in the timetable. 

[Northmoor]

Until recently I was working on development of a software tool for LU that predicts low adhesion down to individual signal sections.  They already use a versions of it on the Northern/Jubilee/Central Lines and DLR, where they can remotely turn down the acceleration or braking rates of trains to compensate for local low adhesion conditions.  It has led to very low levels of SPADs and wheelflats compared to previously. 

It got a lot of interest from main line TOCs and NR as it would have the capability to use the berth-specific broadcast function to drivers in GSM-R; instead of telling them (a) it's October, please drive defensively until January, or (b) there was a low adhesion incident at point A earlier, the driver experienced it three hours ago but he's only just got in off his shift, you could tell drivers to drive normally unless they were told of a specific local risk.  The reduced impact of delays could potentially be huge.

[Steven B]

On 19/01/2022 at 13:55, The Stationmaster said:

So there is really a very simple answer to the problem - cut down the trees and restore railway embankments and cutting sides to the way they used to be which was suitable for a wide range of small mammals and various flora etc - all of which were driven away by tree growth.

 

I don't disagree that fewer trees would lead to less leaves on the line and fewer delays, but a grassed embankment being better for biodiversity than woodland? Really?

 

Some species may do better in open grassland, but mixed woodland supports many more. UK oak trees support 2300 species with over 300 unique to the oak. I'm not sure the same can be said for a patch of ryegrass.

 

Steven B.

 

[The Stationmaster]

1 hour ago, PaulRhB said:

Possibly on some but certainly not widespread on older stock including the 158/9 fleets. As I said above the length of train can make a big difference and it’s not unknown to get an all wheels slide on a 158. 

158s seem to have always been notorious for  for sliding problems and disappearing from track circcuits in pre TV CA days

[The Stationmaster]

7 minutes ago, Steven B said:

 

I don't disagree that fewer trees would lead to less leaves on the line and fewer delays, but a grassed embankment being better for biodiversity than woodland? Really?

 

Some species may do better in open grassland, but mixed woodland supports many more. UK oak trees support 2300 species with over 300 unique to the oak. I'm not sure the same can be said for a patch of ryegrass.

 

Steven B.

 

Can you remember what the lineside was like without its present day jungle and the flora and fauna that inhabited it and posed little or no danger to oassing trains?  I'm all for getting the lineside back to teh way it used to nad getting rid of the overgrown jungles which ruined so much of.  Leaves on the line creating a potential for collisons, trees on the line killing or severely injuring Drivers and injuring passengers - that seems to be a big price to pay just for a few extra trees where there were none previously.  I'm far more in favour of saving a few lives and life changing injuries rather than tolerating  a continued overgrowth of trees on the lineside especially as they are only recent arrivals.

[PaulRhB]

1 hour ago, Northmoor said:

and NR as it would have the capability to use the berth-specific broadcast function to drivers in GSM-R;

Hmmm it’ll have to have some alterations as it currently doesn’t work in first berths as you enter a new area as the trains are slower registering than the broadcast send so they miss it!

one thing I highlighted and eventually they looked into it and confirmed. It’s a pain in our patch as two first berths also protect crossings so an auto broadcast saves a lot of time if there’s a failure.