Failed track maintenance

[Grizz]

2 minutes ago, 96701 said:

I do know that there are discussions at very high level as to how to measure and map the location of railway assets. In the interim some assets have been removed from some mapping systems simply because people found them to be out by some considerable distances. 

 

A lot of people seem to think that it is a simple matter, I can assure them that it is not easy to be absolutely confident that all assets have a known and repeatable location.

Totally. It is an industry within an industry, GPS does help at these days. I was involved in Boots on the Ground asset recording and measuring for RAR (Railtrack Asset Register) on Sussex and some Wessex and Kent back in the day. Wore through a few sets of boots. Mostly done with Red Zone working with Lookout protection, except in platforms etc. It is a complicated and difficult task and is a live system requiring constant updating. ELR (Engineers Line Recognition), TI (Track I.D.) e.g. KJE1 (Keymer Junction to Eastbourne 1).  1100 = UP FAST. 2200 = DN FAST. etc etc. All assets within the boundary fence, between junctions, from the first timber after the last long timber after the crossing to the last timber before the first long timber. This is between to junctions. All measurements are referenced to a particular road. So the down platform at Plumpton is adjacent to the 2200 from X Miles X Chains from Keymer Junction and so on.....

Every asset above, below, adjacent to, on, or in the track. Power supplies, bridges, water mains, cables, everything.

 

It is then broken down and down, in Parent / Child relationships.  Allocated to departments. Mind blowing quantities of info. 
 

 

 

[96701]

5 minutes ago, Grizz said:

Mind blowing quantities of info. 
 

 

 

Also without a known datum. There are examples of signals with 3 known and quantifiable positions, all using different systems and different controlled documents. Admittedly, anybody who cannot find a signal on the railway needs their bumps feeling, but it does demonstrate the difficulties in confirming locations.

[Grizz]

The fact is for manual rectification of track geometry someone still has to go out on the ground to verify where the track geometry fault is, prior to doing the work. The TRT records dynamic track geometry, Top, Line, Twist, Gauge and Offset Structure Gauge etc (Horizontal , Vertical and Cross Level values) against the designs. The person on the ground, if they have been given the correct location, will only be able to check static values if they are undertaking ‘Kango Packing’. Hopefully now most of this is now under mechanised maintenance but not everywhere will justify this.

[rab]

I was quite surprised when I found that this

topic had reached four pages in eight hours.

Having read through it I now see that about

half of it is people playing I know better than

you do.  Oh well I suppose it's kept a few folk

occupied for the afternoon

[Grizz]

3 minutes ago, 96701 said:

Also without a known datum. There are examples of signals with 3 known and quantifiable positions, all using different systems and different controlled documents. Admittedly, anybody who cannot find a signal on the railway needs their bumps feeling, but it does demonstrate the difficulties in confirming locations.

This is a case of poor data verification. The location is either recorded in Metres or Miles Chains Yards and Feet. But getting changed shouldn’t be that difficult. It has a asset number which is based on the NATO numbering system. It’s number is unique and the asset custodian is S&T. It’s location is going to be between point A and Point B. Regardless of what system is used by the maintainer or sub contract maintainer its location on the Network Rail central database should only have one recorded location attributed. This can be verified by someone going out on the ground with a measuring wheel if necessary and checking it. 
 

It will be located on an ELR, with an associated Track I.D. and will be to left, right or above this Track I.D. The datum is based on this. If it has physically moved the the data base would need to be updated.

[96701]

6 minutes ago, Grizz said:

The fact is for manual rectification of track geometry someone still has to go out on the ground to verify where the track geometry fault is, prior to doing the work. The TRT records dynamic track geometry, Top, Line, Twist, Gauge and Offset Structure Gauge etc (Horizontal , Vertical and Cross Level values) against the designs. The person on the ground, if they have been given the correct location, will only be able to check static values if they are undertaking ‘Kango Packing’. Hopefully now most of this is now under mechanised maintenance but not everywhere will justify this.

 

On top of which, if one defect masks another, you don't know whether or not the work that you have done has actually fixed the problem until the next run.

 

[caradoc]

Please be careful commenting on the Stonehaven derailment, the RAIB report will tell the whole story. 

 

[96701]

5 minutes ago, Grizz said:

This is a case of poor data verification. The location is either recorded in Metres or Miles Chains Yards and Feet. But getting changed shouldn’t be that difficult. It has a asset number which is based on the NATO numbering system. It’s number is unique and the asset custodian is S&T. It’s location is going to be between point A and Point B. Regardless of what system is used by the maintainer or sub contract maintainer its location on the Network Rail central database should only have one recorded location attributed. This can be verified by someone going out on the ground with a measuring wheel if necessary and checking it. 
 

It will be located on an ELR, with an associated Track I.D. and will be to left, right or above this Track I.D. The datum is based on this. If it has physically moved the the data base would need to be updated.

This is all very well if you are absolutely sure of your starting point with your wheel. There are real issues where ELRs change at locations where there used to be junctions, but not only no longer exist but have no indication of where they used to be. The mapping systems we use have overlappng ELRs and mileages at such locations because it is not an automatic system (how could it be?).  We have also ended up with negative mileages, long miles and short miles.

 

As I say, it is not simple.

[Trog]

2 hours ago, martin_wynne said:

 

The GPS chips can get a fix several times per second, and can report to the software how reliable it is. The train software could default to counting wheel revolutions between each good fix, if the next fix is not accurate enough to determine distance travelled.

 

My handheld GPS tracker can "see" 20-24 satellites in open country, and reports an accuracy of +/- 1 metre when it does. That's less than the track gauge. It works too -- if I take 2 steps forward the cursor moves on the map, and an arrow appears to show the direction of movement.

 

When under tree cover or other obstructions, the satellite count goes down to typically 7 or 8 satellites visible, and the accuracy drops to around +/- 5 metres -- still less than a 60ft rail length.

 

The tracker cost £300 and uses standard GPS chips. I'm sure Network Rail could run to that.

 

cheers,

 

Martin.

 

I wonder what format the output is in these days? Does the train use GPS to generate a mileage figure? If the gang were then working from the physical mileposts and the one they were using was in the wrong place as a lot of them are could that be the source of the error? If the train reports the defects as co-ordinates were the gang using GPS on an ordinary mobile phone to find them which might be less accurate than Martins super whizzo £300 pound tracker. As according to Goggle Maps on my phone I am not typing this from in front of my fire, but from the middle of my back lawn, and it believed that this chair was in the shed yesterday.

 

Once upon a time the recording train traces used to show AWS magnets as they were something that was easy to automatically pick up and could be used to calibrate the mileage on the trace. As for systematic errors I used to sometimes see the LUL recording train traces, as a result of a year 2000 dodge they all had an error of minus 20 years. The fact that they were in km from the LUL zero datum, rather than miles from Euston also made comparison with our own traces an interesting exercise.

[96701]

45 minutes ago, Trog said:

 

I wonder what format the output is in these days? Does the train use GPS to generate a mileage figure? If the gang were then working from the physical mileposts and the one they were using was in the wrong place as a lot of them are could that be the source of the error? If the train reports the defects as co-ordinates were the gang using GPS on an ordinary mobile phone to find them which might be less accurate than Martins super whizzo £300 pound tracker. As according to Goggle Maps on my phone I am not typing this from in front of my fire, but from the middle of my back lawn, and it believed that this chair was in the shed yesterday.

 

Once upon a time the recording train traces used to show AWS magnets as they were something that was easy to automatically pick up and could be used to calibrate the mileage on the trace. As for systematic errors I used to sometimes see the LUL recording train traces, as a result of a year 2000 dodge they all had an error of minus 20 years. The fact that they were in km from the LUL zero datum, rather than miles from Euston also made comparison with our own traces an interesting exercise.

Trains use Route Setting Tapes which used to come from GEOGIS which is now extinct, however the new system is based on GEOGIS data. People with wheels go out an measure new installations, sometimes using GEOGIS data as a start point, sometimes using a milepost which might not be at an exact geographical mile, after all how accurate were the original measuring techniques, and how many mileages have been re-measured since, hence signalling diagrams showing mileages for signalling equipment which don't natch electrification diagrams which don't match track diagrams, which don't match accurate GPS co-ordinates.

 

A lot of these are controlled documents, so decisions need to made in the light of improved technology and measuring techniques as to how we progress from where we are.

 

AWS magnets can be in 4 places at once dependent on where you start measuring from, which might be where the train starts measuring from, not some spurious buffer stop in London.

[Edwin_m]

6 hours ago, phil-b259 said:

 

GPS is not a navigation system as such - it technically just relates to a system of geostationary satellites from which you can triangulate your position on earth.

 

IF enough satellites are visible to it, GPS will allow a suitably equipped device to position to determine its location in whatever the device manufacturer decides to us as a reference. It could be postcodes, degrees latitude + longitude, number days to walk to Lands end from current position etc.

 

In other words GPS IS 100% accurate and reliable - what manufacturers of receivers may do with the raw digital '1s' and '0s'the receiver churns out a whole other ball game.

 

For trains you need to ask Bombardier, Siemens, Hitachi why their on board passenger information systems don't do that properly - its got nothing to do with GPS itself.

 

The GPS satellites are not geostationary, they are in a relatively low earth orbit that provides global coverage including near the poles where a geostationary satellite would be low on the horizon, and at the equator where they would all be in a line which would reduce the precision.  The system works by three-dimensional triangulation based on measuring the difference between time signals transmitted from each satellite, and if I recall correctly four satellites need to be in view to give a three-dimensional position plus an accurate time.  

 

As probably the first person to put it on a train in the UK, as a BR Research graduate trainee in 1989, I found it to be hundreds of metres out in hilly areas but better than any of the alternatives.  At the time there weren't enough satellites to get a fix at all at certain times, but that and the processing power of the receivers has probably improved it since then.  However it still isn't good enough to distinguish between parallel tracks when mounted on a train, unless enhanced with something called "differential GPS".  

5 hours ago, Derekstuart said:

Didn't we have a mass go-slow not that many years back after the state of rail was exposed? Perhaps that was a turning point.

 

5 hours ago, iands said:

That, I believe, was following the Hatfield crash. 

And as far as I know, no other piece of rail was found that was anywhere near as bad as the one at Hatfield.  The management of the then Railtrack didn't care about engineers, and having missed several opportunities to replace that rail before it failed they didn't have any understanding of whether the same problem might exist elsewhere.  So most trains were severely restricted for several weeks until everywhere had been checked.  

[john new]

At the risk of thread drift mapping distortions are nothing new. Back around the millennium (exact date forgotten) the O Survey began a series of map corrections because of the distortions in the paper maps/traditional surveys that digital surveying and re-mapping was finding. Part of my job at the time required regional liaison on GIS, OS mapping and recording property data so that all agencies had a common base. I clearly recall the delegate from Swindon at one meeting expressing glee and saying something like "At last the original GWR survey data we've always found matched real world dimensions will now also match the O S mapping!"

 

Edited November 16, 2020 by john new

[Derekstuart]

13 hours ago, iands said:

Hi Derek,

Afraid I can't answer your specific question(s), but if you have a concern about your observations regarding the cross-over at Colchester, did you try to report your concern to Network Rail or ORR?

 

13 hours ago, phil-b259 said:

 

I have no idea what the p-way standards are for this.

 

However if you are concerned, best course of action is to take a photo, highlight the area of concern, then e-mail it in with as much information as the where it is (things like road bridges can be useful as they should all have an ID plate stating the Engineers line reference (ELR) and a mileage on them. you can then state something like "defect is 50m London side of Bridge XYZ on line ABC")

 

It can then be passed to the appropriate department (It works because I have seen items on the p-way fault list like 'loud bangs heard by MOP as trains go past location X)

 

If you feel that it is being ignored then keep bothering NR - like any large organisation its not perfect but persistence pays off.

 

See https://www.networkrail.co.uk/communities/contact-us/

https://communications-crm.custhelp.com/app/ask?var=incident

 

 

Thank you.

 

I should have stated this was from around August 2017 to January 2018 before I get the people responsible for it in 2020 into trouble. It was spotted from the train as we sat at the down fast platform. I have  photos somewhere.

 

I reported it to the 'guard' (or whatever they are called now) and he said he'd report it to whomever he reports to. The guard from Jan 18 was certainly a real railway professional and I am 100% certain he'd have reported it. 

 

As a point of note, this isn't the only time I've had a concern on the GEML. One time out of Norwich, just before the bridge on a crossing, it felt as if we'd dropped in the dirt on an axle (I'm not saying that's what happened, but how similar to how I imagine it must feel). It stopped suddenly when we went over a crossing not long after. Again this was reported to the guard (the same one as in Jan 2018 and I know he reported it as he told me the next time I saw him. He mentioned he'd asked for a follow up and told there was no fault found- but at least it was checked).

Sorry for the off-topic but I wanted to clarify the historic rather than contemporary nature of that observation.

[Oldddudders]

18 hours ago, iands said:

That, I believe, was following the Hatfield crash. 

And that accident caused the entire industry to go into meltdown. It was like Clapham in the enormous after-shock, but this time trains went slowly, timetables became unviable and the service virtually collapsed. Spurred on by a gleeful popular press, the public lost faith in train travel. I was told that Virgin's income dropped so low they hadn't the cash to pay wages, let alone train-leases or charges to Railtrack etc. National Express, for whom I then worked, amalgamated three franchises into one at TOC HQ level to save costs. 

 

I had previously worked closely with one of the people who ended up in the dock in the wake of Hatfield. A Chartered Engineer, he had not impressed me, or others, most of whom were very senior engineers, in that role. 

[alastairq]

15 hours ago, phil-b259 said:

Not so!

 

You can the perfect driver yourself yet be driven into by someone who is drunk, under the influence of drugs, evading the police, texting on their phone, arguing with their co-occupants, has a mechanical failure, is driving an unroadworthy vehicle, etc.

Etc etc

 

You're still missing the point entirely.

 The safety of train or air travel is down to a number of individuals.....driver or pilot, maintenance of vehicle and infrastructure, etc.

The objective being, the safety of a passenger. That being entirely in the hands of others. With the passenger being passive, IE, having no input into that safety or wellbeing.

That is not the same as road travel in individual cars! In which our 'passenger' suddenly takes responsibility  for their own safety and welfare, by being the  driver.

To equate rail and air transport with roads, one can only compare with bus travel. Where again, our 'passenger' places their safety and wellbeing in the hands of ONE or more different  individuals [or several in one includes maintenance, etc.] That one individual [the bus driver] is responsible for protecting the passenger from all threats to safety and welfare, in the same manner as the train driver or aircraft pilot.

 

Nothing whatsoever to do with driving standards, threats or evils. Or whatever else you perceive as nastiness out there on the roads.

 

As for investigations?

Road traffic Incidents are indeed, investigated.

Either by the Constabulary...to ascertain facts and events. Or by insurance companies, to ascertain liability.

The DVSA correlate issues to identify emerging levels of risk, etc, so that the driver training syllabuses can be altered or amended to minimise new perceived risks.

The Highway Code can, and is, adjusted to suit emerging trends on the roads.....[and they welcome input from individual road users, especially from the driver education industry]

Individual drivers can be re-educated. Especially vocational drivers. 

Not forgetting the input from, and to, the road transport industry itself.

 

 

 

 

[iands]

17 minutes ago, Oldddudders said:

And that accident caused the entire industry to go into meltdown. It was like Clapham in the enormous after-shock, but this time trains went slowly, timetables became unviable and the service virtually collapsed. Spurred on by a gleeful popular press, the public lost faith in train travel. I was told that Virgin's income dropped so low they hadn't the cash to pay wages, let alone train-leases or charges to Railtrack etc. National Express, for whom I then worked, amalgamated three franchises into one at TOC HQ level to save costs. 

 

I had previously worked closely with one of the people who ended up in the dock in the wake of Hatfield. A Chartered Engineer, he had not impressed me, or others, most of whom were very senior engineers, in that role. 

I was with Railtrack LNE at the time, and was involved with the immediate aftermath of Ais Gill, Hatfield, Potters Bar and Great Heck. Not exactly a happy time for all concerned.

[62613]

19 hours ago, phil-b259 said:

 

What I am telling you is that the answer to your alarmist question "is the railway safe" is yes it is safe to use.

 

That does not mean everything is perfect - I have never said it is. There are many things which require improvement, but that is not the same thing as saying the railways are inherently dangerous to use (as you are implying with your question) - particularly when the alternative to the train  for 99% of people is going to be getting into a motor car and being far more likely to be involved in a serious accident on our road network!

When you think about it, railways are extremely dangerous; that's why we have all the various rules and regulations for running them.

[alastairq]

 Not forgetting, the road network is vastly greater in mileage, and user numbers, than the rail network.

 

Sounds like a bit of a ''London-centric'' argument?    

It would be interesting to compare injury rates per journey undertaken. Especially on the roads?

Because quoting the threat to life & limb on our roads sounds rather sensationalist to my ears.

Certainly, given the numbers of hours I personally used to spend out on the roads, not my experience  either.

Edited November 17, 2020 by alastairq

[Phil Parker]

1 hour ago, alastairq said:

The Highway Code can, and is, adjusted to suit emerging trends on the roads.....

 

Adjust it all you like. Most drivers never read a copy once they pass their test.

 

20 minutes ago, 62613 said:

When you think about it, railways are extremely dangerous; that's why we have all the various rules and regulations for running them.

 

Everything is dangerous. Trains generally can't steer themselves into the path of other trains, unlike road transport.

 

I've often thought that if you invented cars today, the idea that opposing traffic with closing speeds of 120mph, could be kept apart by painting a line along the middle of the road, would be laughed at. I suspect that mandatory re-examination would also be included in any plan, as it is for airlines and I think rail. Suggest that every driver has a refresher twice a decade and watch for the "war on the motorist" headlines.

[Talltim]

The 2018 rate for road fatalities was 5.38 per billion vehicle miles. (1,784 total, 331.5 billion vehicle miles)  Difficult to translate that for rail as a vehicle carries far more occupants (and road vehicles can obviously carry a varying number of occupants too). Also road figures include freight, whereas rail ones don't

However https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwjTk5_xtontAhVKQRUIHVfoASwQFjABegQIBhAC&url=https%3A%2F%2Fdataportal.orr.gov.uk%2Fmedia%2F1452%2Frail-safety-statistics-2018-19.pdf&usg=AOvVaw1w_hcBGjeg2GCOuxskub_E

Quote

The number of passenger fatalities increased to 17 in 2018-19 - an increase from 9 in 2017-18. This is the highest number in the last ten years.
There were no passenger fatalities as a result of a train accident in 2018-19.

Obviously this figure is only those who have chosen to travel by train, as opposed to those killed on the railway (348 for 2018-19). I would assume the road figures only include those who have died in some sort of vehicular collision, rather than, for example dying by jumping off a bridge onto a road or while working on the road infrastructure.

Edited November 17, 2020 by Talltim

[Michael Hodgson]

21 hours ago, phil-b259 said:

 

 

Whitewash doesn't tell you much more than the track is a bit bumpy - it cannot give you a scientific measurement which can then be used to determine the severity of the defect.

 

 

It can also tell you if somebody's been nicking coal from your fuel stockpile

[highpeakman]

22 hours ago, phil-b259 said:

 

GPS is not a navigation system as such - it technically just relates to a system of geostationary satellites from which you can triangulate your position on earth.

 

IF enough satellites are visible to it, GPS will allow a suitably equipped device to position to determine its location in whatever the device manufacturer decides to us as a reference. It could be postcodes, degrees latitude + longitude, number days to walk to Lands end from current position etc.

 

In other words GPS IS 100% accurate and reliable - what manufacturers of receivers may do with the raw digital '1s' and '0s'the receiver churns out a whole other ball game.

 

For trains you need to ask Bombardier, Siemens, Hitachi why their on board passenger information systems don't do that properly - its got nothing to do with GPS itself.

 

GPS satellites are not geostationary. The constellation of 24 satellites constantly move in orbit around the earth - all with different orbits so that there will always be a number in view at any time. There are also two types of signal transmitted, one of which is used by the military and one, slightly less accurate and more susceptible to jamming, for public use. As you would expect the military grade signal is not available for public use but it is the prime reason for the satellites being there. Accuracy of position is determined by triangulation based on measuring the time taken for the satellites signals to reach the receiver. Accuracy of position information will be determined by quality of software, and the equipment used plus getting a clear view for receiving the signals. In cities where large tall buildings (which both shield the signal but also reflect it causing errors), trees and tunnels occur there can be temporary loss of sight of satellites which a reduction in accuracy of position although not usually major. Usual accuracy should be, at worst case, 30 metres but is usually better. Sometimes maintenance (or fault) of satellites may reduce accuracy for short periods but, again, usually not major as with the constellation constantly moving new satellites come up over the horizon continuously. Usually the receiver has a capability of smoothing out some of these perturbations. The faster you are moving the greater the potential for losing positional information.

Some of this can be seen when using a car sat nav in major cities. I have not had it happen in the UK but in Toronto centre I once got lost because the sat nav just lost all ideas of where it was in amongst the sky scrapers. 

Edited November 17, 2020 by highpeakman

[Trog]

15 hours ago, 96701 said:

Trains use Route Setting Tapes which used to come from GEOGIS which is now extinct, however the new system is based on GEOGIS data. .....................

 

That might explain a lot as myself I found that most of the data in GEOGIS particularly for the age and type of the track which was what I had the most to do with was so unreliable as to not be worth reading.

 

GEOGIS was an old system where everything used obscure letter and number codes to reduce the size of the data. To be able to enter the data you needed to know what you were talking about and what each code really referred to, however GEOGIS input was often delegated to the temp, or most junior member of staff. The way the input screen was set up did not help either as for example the 1950's 113lb FB rail and normal grade steel were offered before the currently used 113A FB and wear resistant A grade steel. So those of little knowledge just typed in 113lb FB normal grade as the newly installed rail not knowing that such a thing had not been rolled for over fifty years. According to GEOGIS there was mile upon mile of 113lb FB rail on the area I was interested in, but on track I only ever found a single 60'-0" rail, and I looked at and recorded the track details for years on end.

 

The line designations were also denoted by a four digit number, get that wrong and they often did and you buggered up the data for two lines.

 

Mileages are often dubious with real life not matching what is recorded in the Sectional Appendix, there was one junction were the ELR changed and theoretically there was a gap between two red zone working prohibited areas, right at the most dangerous part of the junction. As the RZP areas matched the Sectional Appendix while in reality there was another 60'-0" of track right on the switch diamond that had not been noticed. I assume the junction having moved since the first Sectional Appendix was produced.

 

This was not really a problem until towards the end of my time at Network Rail as I would just go out to site and pick up the details I needed from real life. But as the safety nutters started to insist that all work be done from your desk using records instead, the fact that the records were rubbish started to become more worrying.

[caradoc]

7 hours ago, Derekstuart said:

As a point of note, this isn't the only time I've had a concern on the GEML. One time out of Norwich, just before the bridge on a crossing, it felt as if we'd dropped in the dirt on an axle (I'm not saying that's what happened, but how similar to how I imagine it must feel). It stopped suddenly when we went over a crossing not long after. Again this was reported to the guard (the same one as in Jan 2018 and I know he reported it as he told me the next time I saw him. He mentioned he'd asked for a follow up and told there was no fault found- but at least it was checked).

 

Following the Potters Bar accident Network Rail instituted a formal procedure for dealing with infrastructure faults reported by members of the public, to ensure that these did not get dismissed or ignored.

 

3 hours ago, Oldddudders said:

And that accident caused the entire industry to go into meltdown. It was like Clapham in the enormous after-shock, but this time trains went slowly, timetables became unviable and the service virtually collapsed.

 

Absolutely; Soon after Hatfield the Route Director in Scotland decreed that our part of the WCML would be closed at 0800 the next day to allow the route to be checked for Gauge Corner Cracking. However not everybody outside Scotland seemed to have been told, so it was left to me as nightshift Senior Controller to tell Carlisle SC to stop sending trains towards Lockerbie, and to deal with the fallout from them and the Train Operators. 

 

 

[Derekstuart]

4 hours ago, highpeakman said:

There are also two types of signal transmitted, one of which is used by the military and one, slightly less accurate and more susceptible to jamming, for public use. As you would expect the military grade signal is not available for public use but it is the prime reason for the satellites being there.

 

Hello HPM.

Are you sure about this? My understanding is that the public restrictions were removed some time back. Originally you could only navigate to around 90ft accuracy, whereas now it can be 'up to' 3ft.

Naturally, the US Department of Defense, who own the system, have the ability to remove or restrict this public feature at any time. I believe their reason for removing this restriction in the first place was because it served no purpose. In a time of war they can restrict it if they need to.

 

However, I stand to be corrected on this if I am wrong.