Jump to content
 

RAIB report : Bognor derailment


Recommended Posts

  • RMweb Gold
1 minute ago, bigherb said:

The report said the indicators where not illuminated, is this another case of asking someone to notice something not there as opposed to a positive indication?

 

True. In a panel-box, the independent point switches would be flashing a light to show "out of correspondence" rather than both N and R indications just being blank. But as Phil says, with an indicator provided just above the lever concerned - and of course the lever has been shortened to show it is not mechanical - it should be automatic for the signaller to watch the lights change after pulling. 

Link to post
Share on other sites

  • RMweb Premium
2 minutes ago, bigherb said:

The report said the indicators where not illuminated, is this another case of asking someone to notice something not there as opposed to a positive indication?

 

 

Its a function of these type of indicators and the fact that we only provide an electrical feed from the points when they are fully detected in normal / reverse

 

The traditional 'moving needle' brass type had three positions (Normal to the left, Reverse to the right and 'wrong' at the bottom) and if the points were not detected then the needle would simply hang vertically under the influence of gravity. Unfortunately these indicators haven't been in production since the 1950s....

 

On a signalling panel setup additional circuitry is provided to have 3 lights above / below the switch showing Normal, Out Of Correspondence (i.e. we have no idea what the points are doing) and Reverse. I presume IECC setups have something similar.

 

Incidentally this later setup is far from foolproof - only last month a signaller working the Redhill panel managed to send a train through a set of points left in the wrong position due to cautioning being needed after a fox had gone and chewed a cable to a signal head. That was a fun night shift trying to fix the resultant damage.......NOT

 

 

 

 

 

  • Interesting/Thought-provoking 1
Link to post
Share on other sites

20 hours ago, 101 said:

It's a complete comedy of errors - a signalman setting the route by pulling levers according to a crib sheet, then when he can't pull the signal off he doesn't question why.

Then when he can't pull a second one off and doesn't question that either

 

Then two drivers are talked past signals but both - unfortunately from an ex drivers point of view - manage to run through / out of catch points 

 

If the first driver had stopped to question why the wrong route was set as he should have,  perhaps it all could have been avoided

It's known as the ducks lining up. Fortunately the RAIB don't appear to be doing what many here are doing which is to apportion blame. If you're tempted to say "He ought to have known that....." it's time to look further.

Human beings always have and  always will make mistakes and sometimes fail to notice things that in retrospect seem totally obvious. A system that relies on human beings not making mistakes is an accident looking for somewhen to happen and the whole history of railway safety (it applies to air safety and other areas as well) has been the gradul and painful learning of how to avoid that. 

I'm no expert on signalling but I did study ergonomics as part of my engineering degree and have been responsible for hazard assessments in my own industry. (Sometimes the hazards that our assessments picked up were those we'd never thought of before we started carrying them out because they seemed so obvious)

 

From that outside perspective,  there are several things here that seem worthy of note.

Why does the failure of a battery perfoming a function as vital as powering point motors not throw up an alert?

More importantly. Why, when motorised points are operated from a mechanical frame, is there not more indication of that fundamental difference than a label attached to an adjoining, used to be an FPL, locking lever. An operator may "know" that certain levers have a different function but, in the absence of sufficient cues, is very likely to overlook that.  Are all the points in that frame accompanied by detection indicators or only some of them and how closely are the detections associated with the relevant point lever? 

 

The fact that the RAIB are citing so many parallel incidents also seems significant. 

Edited by Pacific231G
  • Like 2
Link to post
Share on other sites

  • RMweb Premium
1 hour ago, Pacific231G said:

 

From that outside perspective,  there are several things here that seem worthy of note.

Why does the failure of a battery perfoming a function as vital as powering point motors not throw up an alert?

 

 

Point batteries don't tend to fail outright in the sense that they will continue to have 120V sitting across them even when failed. The issue normally is that if several of the cells have degraded / leaked / etc then there will not be enough current to power the points across (a point machine will need around 7A or so to go over and if it hits an obstruction currents as high as 15Amps can be drawn without blowing the 20A supply fuse) , particularity if several are trying to move at once.

 

As such the the first indication there is a problem is when points fail to move across completely - regular checks by techs will of course pick up things like cells starting to split or degrade, but this in itself doesn't give an indication as to whether the batteries will fail that day, next week, next month, etc.

 

1 hour ago, Pacific231G said:

More importantly. Why, when motorised points are operated from a mechanical frame, is there not more indication of that fundamental difference than a label attached to an adjoining, used to be an FPL, locking lever.

 

 

As the RAIB highlight, the alterations the signalling standards require to the lever colours and identification plates were not carried out when the points were motorised. This is a failing of the project (which was to motorise all the points at Bognor as part of a life extension / reliability improvement scheme a few years ago now) and is an own goal as far as NR is concerned.

 

1 hour ago, Pacific231G said:

 Are all the points in that frame accompanied by detection indicators or only some of them and how closely are the detections associated with the relevant point lever? 

 

 

ALL motorised points are fitted with indications for the very simple reason that the physical feedback provided by the lever is gone! On a mechanical set of points then if there is an obstruction you will not get the lever fully over while if the point rodding breaks then the lever will become ridiculously light / easy to pull.

 

The point indicators are fitted directly above the point lever concerned - so as Oldudders says its easily visible as the signalman operates that lever.

Edited by phil-b259
  • Informative/Useful 1
Link to post
Share on other sites

  • RMweb Gold

On a workstation or panel the signaller is provided with (and required to use) route cards (which are, I believe, the property of the Signalling Engineer). These specify the positions that points are to be placed in (for all signalled routes) when a signal is to be passed at danger. Route cards are not normally provided for lever boxes though, although one large frame I worked was indeed supplied with them. The report mentions an informal guide to lever movements which is not likely to be the same thing. It has long been my belief that every box with pointwork should be issued with route cards. Had this been the case, the signaller would probably have set the route according to the relevant card and checked that the points indications were correct, at which point the fact that points were out of correspondence would have been obvious and movements could not be made until the points were secured.

  • Like 1
  • Agree 1
Link to post
Share on other sites

7 hours ago, Oldddudders said:

True. In a panel-box, the independent point switches would be flashing a light to show "out of correspondence" rather than both N and R indications just being blank. But as Phil says, with an indicator provided just above the lever concerned - and of course the lever has been shortened to show it is not mechanical - it should be automatic for the signaller to watch the lights change after pulling. 

It wasn't clear from the images in the report that the motorised point levers had been shortened and were under the inficators. That makes the human interface* rather better than it at first appeared.

7 hours ago, phil-b259 said:

The traditional 'moving needle' brass type had three positions (Normal to the left, Reverse to the right and 'wrong' at the bottom) and if the points were not detected then the needle would simply hang vertically under the influence of gravity. Unfortunately these indicators haven't been in production since the 1950s....

 

On a signalling panel setup additional circuitry is provided to have 3 lights above / below the switch showing Normal, Out Of Correspondence (i.e. we have no idea what the points are doing) and Reverse. I presume IECC setups have something similar.

 

Incidentally this later setup is far from foolproof - only last month a signaller working the Redhill panel managed to send a train through a set of points left in the wrong position due to cautioning being needed after a fox had gone and chewed a cable to a signal head. That was a fun night shift trying to fix the resultant damage.......NOT

 

No system is foolproof- foolishness just get more cunning- but I agree about the preference for analogue indicators. Designing systems that accord with the way humans naturally interact with the world rather than what is easiest to provide is definitely preferable.

 

*I'm very conscious of a particular light aircraft (often a fruitful soure of examples of poor ergonomics) where the throttle, carbutettor heat and mixture controls  were neatly arranged as a row of three identical levers only differentiated by colour. Pulling the mixture control by mistake, usually during a high workload phase such as aproach and landing, would stop the engine with potentially fatal results.

Link to post
Share on other sites

  • RMweb Gold
1 hour ago, Pacific231G said:

It wasn't clear from the images in the report that the motorised point levers had been shortened and were under the inficators. That makes the human interface* rather better than it at first appeared.

It has been standard practice for many years to shorten levers when they operate motor points, colour-light signals or other non-mechanical linkages. An early H&S thing, really, because signalmen pride themselves on having the knack for getting the up distant off or 37 points over first time. The effort they put into such operations is considerable, and were they to suddenly to find that the lever they are pulling offers no such physical resistance, being hurled backwards across the box with a high risk of injury is a likely outcome. 

  • Agree 1
Link to post
Share on other sites

One thing that seems to have been overlooked in this discussion and could well be contributory is that in a mechanical box you cannot move the blue point lock lever unless the point is fully over.  This means that if the levers are in the correct positions, then that is the proof that the route is set correctly*.  However in this case they only operated the interlocking in the frame and were no longer attached to the points.  They should have had the top half painted white and the signal diagram updated to reflect this, but this had not happened.  There was a small tape label added to the description on the lever, but if you are trained that you can only move the blue lever when the points are set unless the top half is painted white, then it is not unreasonable that he took the position of the levers to signify the points were in the correct position, as according to the signal diagram and the way they were painted this should have been the case.

 

*There is one freak accident that I know of that this was briefly not the case in a mechanical box due to an extraordinary set of circumstances where I believe the signalman in that instance was completely blameless.

Edited by Titan
  • Agree 2
Link to post
Share on other sites

8 minutes ago, Titan said:

One thing that seems to have been overlooked in this discussion and could well be contributory is that in a mechanical box you cannot move the blue point lock lever unless the point is fully over.  This means that if the levers are in the correct positions, then that is the proof that the route is set correctly.  However in this case they only operated the interlocking in the frame and were no longer attached to the points.  They should have had the top half painted white and the signal diagram updated to reflect this, but this had not happened.  There was a small tape label added to the description on the lever, but if you are trained that you can only move the blue lever when the points are set unless the top half is painted white, then it is not unreasonable that he took the position of the levers to signify the points were in the correct position, as according to the signal diagram and the way they were painted this should have been the case.  

Whilst that is true, and the top half of the lever certainly should have been white, it would have been immediately obvious on actually pulling the lever that you were only working the interlocking and not a physical fpl. In my experience, fpls were usually a harder actual pull than the points themselves, certainly no easier, and that is very different from the feel of a lever retained just to work the interlocking.

I was always taught to watch the indicator lights on "electric" levers, it only needs a couple of momentary glances and quickly becomes second nature and I have noticed that experienced bobbies, even old hands, always did it too. I find it particularly surprising that a relief man seemingly didn't do it as a matter of course, as it would have provided familiar confirmation in a relatively unfamiliar box.

Link to post
Share on other sites

  • RMweb Premium
On 19/12/2020 at 21:41, Titan said:

One thing that seems to have been overlooked in this discussion and could well be contributory is that in a mechanical box you cannot move the blue point lock lever unless the point is fully over.  This means that if the levers are in the correct positions, then that is the proof that the route is set correctly*.  However in this case they only operated the interlocking in the frame and were no longer attached to the points.  They should have had the top half painted white and the signal diagram updated to reflect this, but this had not happened.  There was a small tape label added to the description on the lever, but if you are trained that you can only move the blue lever when the points are set unless the top half is painted white, then it is not unreasonable that he took the position of the levers to signify the points were in the correct position, as according to the signal diagram and the way they were painted this should have been the case.

 

*There is one freak accident that I know of that this was briefly not the case in a mechanical box due to an extraordinary set of circumstances where I believe the signalman in that instance was completely blameless.

That was the first thought that occurred to me. The signalman (new, unfamiliar with the box and tired, and a person for whom I have some sympathy) was not relying on the position of the black points levers, but on the position of the blue FPL levers.

 

Should they still be painted blue? Painting them a different colour would be  simple change to make.

On 19/12/2020 at 22:02, bécasse said:

Whilst that is true, and the top half of the lever certainly should have been white, it would have been immediately obvious on actually pulling the lever that you were only working the interlocking and not a physical fpl. In my experience, fpls were usually a harder actual pull than the points themselves, certainly no easier, and that is very different from the feel of a lever retained just to work the interlocking.

I was always taught to watch the indicator lights on "electric" levers, it only needs a couple of momentary glances and quickly becomes second nature and I have noticed that experienced bobbies, even old hands, always did it too. I find it particularly surprising that a relief man seemingly didn't do it as a matter of course, as it would have provided familiar confirmation in a relatively unfamiliar box.

FPLs with locking bars were hard to pull, but Bognor Regis (and probably all the boxes the signalman had ever worked in) were track circuited, so FPLs are probably easy to operate. Remember that the point levers will also have been easy to pull, so going by comparison of one lever to the next, nothing may have seemed amiss, and perhaps the signalman merely thought as he pulled the levers that everything was immaculately maintained.

Link to post
Share on other sites

Quote

That was the first thought that occurred to me. The signalman (new, unfamiliar with the box and tired, and a person for whom I have some sympathy) was not relying on the position of the black points levers, but on the position of the blue FPL levers.

There is no practical difference which lever he relied on, he should have checked the indicator. However this situation was greatly helped by the cheapskate job done by NR when converting the points to electric. Apart from the minor issue of correctly painting the levers a proper job would have fitted to point levers with electric check locks so that the lever movement could not be completed, and hence the locking released unless the detection was recieved correctly. When I learnt my trade check locks were a fundamental part of operating a point motor from a mechanical interlocking frame. It was probably considered that having the detection in the signal aspect circuit was good enough, clearly it wasn't.

  • Agree 3
  • Informative/Useful 1
Link to post
Share on other sites

1 hour ago, Jeremy C said:

 

FPLs with locking bars were hard to pull, but Bognor Regis (and probably all the boxes the signalman had ever worked in) were track circuited, so FPLs are probably easy to operate. Remember that the point levers will also have been easy to pull, so going by comparison of one lever to the next, nothing may have seemed amiss, and perhaps the signalman merely thought as he pulled the levers that everything was immaculately maintained.

FPL levers where TCs provide the approach locking are clearly much easier to pull than those where a locking bar is fitted, but I still think that any experienced bobby (and almost by definition a relief man should be experienced) would instinctively feel the difference between a lever that was operating point rodding plus a physical fpl and one that was just operating the requisite locking in the frame.

Link to post
Share on other sites

  • RMweb Gold
7 minutes ago, bécasse said:

any experienced bobby (and almost by definition a relief man should be experienced) 

Welcome to 2020. This man joined the railway in July 2018, and by Feb 2020 was on the relief. Things are just not the way they were. 

  • Agree 1
Link to post
Share on other sites

14 minutes ago, Oldddudders said:

Welcome to 2020. This man joined the railway in July 2018, and by Feb 2020 was on the relief. Things are just not the way they were. 

Even though things are obviously just not the way they were, I would still expect a bobby to pick up enough experience in 20 months, especially with some time as a relief, to instinctively realise the difference. I have only ever worked frames by invitation but my brain would have instantly told me that something unexpected was happening - in fact I suspect that my first reaction to an easy pull would have been to query whether I was pulling the right lever. I was always particularly careful pulling fpls, especially given the Southern's habit of working two of them off the same lever which could make for some very hard pulls with a curved layout.

Link to post
Share on other sites

  • RMweb Premium
2 hours ago, 97xx said:

Question: sorry if I missed this, but are the position indicators powered by the same (defective) battery supply?

They are lamps, N and R, and since the points were wrong, neither were lit. This, of course, means that they fail safe, but if the signalman was not looking for an indication, the absence of one would not alert him that anything was wrong. Of course he should have known of the indicators and looked at them, particularly when authorising a movement past a signal at danger.

 

The report shows an indicator lit for another set of points, presumably taken shortly after the accident, and there is no suggestion in the report that the indicators weren't working normally.

Link to post
Share on other sites

  • RMweb Gold

There seems to have been a tendency in this case to assume a signal had failed if it couldn't be cleared—rather than looking at any other possibility. While fatigue and time pressure may have helped lead to this, I wonder if signals not clearing due to a failure of the signal itself are really that common?

Link to post
Share on other sites

On 19/12/2020 at 11:15, phil-b259 said:

On a signalling panel setup additional circuitry is provided to have 3 lights above / below the switch showing Normal, Out Of Correspondence (i.e. we have no idea what the points are doing) and Reverse. I presume IECC setups have something similar.

Just to complete this angle, Out of Correspondence is indicated when the points are not detected in either position, or when the position the are detected is different from that which they have been controlled to (either by setting a route or by operating the point switch).  IECCs show this by flashing the screen character representing that set of points - I think (some?) panels flash the appropriate lights on the track diagram.  

 

One thought that occurs with indicators is that they are aligned with but some distance behind the lever they relate to, especially when the lever is reversed.  If the signaller was to one side when operating the lever there might be a risk of mis-reading the indicator for an adjacent lever.  

 

Edited by Edwin_m
Link to post
Share on other sites

  • RMweb Gold
21 hours ago, bécasse said:

I would still expect a bobby to pick up enough experience in 20 months

 

22 hours ago, bécasse said:

but I still think that any experienced bobby (and almost by definition a relief man should be experienced)

It’s just a job you can apply for at any stage in your career, including off the street, with a flexibility premium that makes it very attractive for some. 

Link to post
Share on other sites

  • RMweb Premium
2 hours ago, D9020 Nimbus said:

There seems to have been a tendency in this case to assume a signal had failed if it couldn't be cleared—rather than looking at any other possibility. While fatigue and time pressure may have helped lead to this, I wonder if signals not clearing due to a failure of the signal itself are really that common?

 

Signals themselves rarely 'fail' as such - when signals are said to have 'failed' the actual reason is usually the interlocking or other kit has a problem and it protects safety by failing to let the signal change to a proceed as a result.

 

If fitted, TPWS can easily cause a 'signal failure' that isn't a signal failure as because TPWS is not inherently 'fail safe' - if its powered down when it should be active we make the interlocking behave as if the signal itself is blank.

 

Another 'signal failure' thats not a signal failure might be when the signallers emergency replacement is activated (well technically deactivated as its has to be 'fail safe') due to a lineside cable fault and the necessary feed from the signal box is not being maintained.

 

The bigger the area controlled by the interlocking and the more equipment connected to it the grater the possibility of something going wrong. In a relay interlocking for example it might be that the points have moved correctly on site but a single high resistance electrical contact (e.g. 500 ohms rather than 0.5 ohms) inside the relay room causes the interlocking to think otherwise even though the relay insides are physically in the correct position. The rapid advancement in logger technology and remote condition monitoring from site makes a big difference here as the S&T techs can quickly focus their attention on the right bit of the overall signalling system.

 

In principle signalling is not hard - its fairly easy to come up with a basic list of things you want to check before letting a train proceed (i.e. All points must be set right, no conflicting routes set, prove the section between signals isn't already occupied by a train, make sure a train driver never gets to a blank signal....etc). The difficult bit is devising a system makes sure all those are implemented before the signal changes to a proceed aspect in metal or electrical form, and perhaps more importantly putting in back the feedback loops which will see the signal revert to danger if any of the things on the above list cease to still be true.

Edited by phil-b259
  • Informative/Useful 1
Link to post
Share on other sites

  • RMweb Gold
58 minutes ago, PaulRhB said:

 

It’s just a job you can apply for at any stage in your career, including off the street, with a flexibility premium that makes it very attractive for some. 

It would appear that that is the railway in 2020. Just a job. Like cleaning windows or being cashier at a petrol station. Long gone are the days of school-leaver to retirement, which I enjoyed. If properly managed, the railway is still safe. But the chasm of understanding by the signaller in this report suggests it is sometimes otherwise. 

  • Agree 4
Link to post
Share on other sites

  • RMweb Gold

Definitely one of those situations where the factors all added up - the point motorizing being 'done on the cheap' with the levers not repainted and the locking not adjusted, then the signalman being unfamiliar with the system and being fatigued, and then the actual failure itself of the batteries.

 

There's particular parallells with Knaresborough, where a set of power points hadn't gone fully over, and the MOM covering the relief turn in the box authorised the train to pass the inner home at danger, leading to it splitting the points and derailing - again, the lever could be reversed as the locking was satisfied, but the signal wouldn't clear because there was no detection.

 

It's something that was repeatedly drummed into me as part of my training at the MHR - always check your repeaters!

  • Like 1
Link to post
Share on other sites

1 hour ago, Nick C said:

Definitely one of those situations where the factors all added up - the point motorizing being 'done on the cheap' with the levers not repainted and the locking not adjusted, then the signalman being unfamiliar with the system and being fatigued, and then the actual failure itself of the batteries.

 

There's particular parallells with Knaresborough, where a set of power points hadn't gone fully over, and the MOM covering the relief turn in the box authorised the train to pass the inner home at danger, leading to it splitting the points and derailing - again, the lever could be reversed as the locking was satisfied, but the signal wouldn't clear because there was no detection.

 

It's something that was repeatedly drummed into me as part of my training at the MHR - always check your repeaters!

The possibillty occurs to me that volunteers on heritage railways may actually be more thoroughly trained- or at least spend more time learning specific installations from more experienced colleagues -  than their professional counterparts. What we don't know of course is what other boxes this particular signaller was regularly operating and whether the Bognor Regis box had characteristics that were different from the boxes he was more regualrly operating;  were they perhaps all mechanical?

Edited by Pacific231G
  • Interesting/Thought-provoking 1
Link to post
Share on other sites

1 hour ago, Nick C said:

Definitely one of those situations where the factors all added up - the point motorizing being 'done on the cheap' with the levers not repainted and the locking not adjusted, then the signalman being unfamiliar with the system and being fatigued, and then the actual failure itself of the batteries.

Going by the picture in the report, the levers looked freshly painted - just not in the correct colour scheme!  

Link to post
Share on other sites

I would still be interested to know whether signalmen at Bognor regularly had to provide authority to pass at danger, or at the other extreme whether it was the first time this signalman had ever had to do so?

Edited by JimC
  • Agree 1
Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...