Cost Vs Safety

[G0ORC]

When I left the industry a few years ago there was an accepted way of grading risk in terms of cost, time and deaths.

i.e. if a safety project would cost, say £10 million to implement system wide in order to save 1 death every 10 years then it was considered not value for money.

I don't know what the current measure is in these terms but you can bet there is one.

 

[cjmillsnun]

Everywhere tbvfh.

But trying to be more mindful and in the interest of debate.

'Stopping Short'
Why aren't trains fitted with gps and technology to prevent door openings when short of the stopping point ?
What about other tech ?
Why cant all trains stop at a single point ?

1.Some are being retrofitted with this technology
2.What kind of other tech?
3. Different trains are of different lengths. People tend to wait in the middle of a platform, therefore to shorten dwell times, trains are stopped at different points relative to their lengths.

'Level Crossings'
Remove them all and use bridges, tunnels, viaducts etc.

and people will still cross the tracks where they shouldn't. At least with a level crossing you have a defined (and in most cases monitored and interlocked with signalling) crossing point.

SPAD's
Can trains be stopped BEFORE the signal instead of within the overlap ?
Is ERTMS the solution ?
Whats the best train/signal protection (TPWS/ATP/ATO etc ?)

1. Yes. ATP does this, as it has indication of the signal ahead and has an intervention braking profile that will stop the train before the danger signal.
2. ERTMS is certainly a solution.
3. It depends. If you believe a human is better being in charge, then ATP, if you want the train to drive itself then ATO. TPWS is a cheaper alternative to ATP that in most cases will do the job perfectly safely.

I emboldened the Central Door locking because its a system that's been put in place to prevent the casual door opening of the past. Rolling stock has made advances in train safety. Its still moving at a snails pace though with rolling stock being passed around the country rather than shiny new super safe and comfortable trains with all the bells and whistles...

CDL was actually a very cheap way of improving safety and was added due to incidents of doors opening.

The biggest problem is not reducing the risk, but who is going to pay to reduce said risk. In every case, that is the passenger, either directly through their rail fare or through taxation and government subsidy. I'm sure you don't want to pay more for your ticket or pay more tax!

 

[GB]

It should also be noted that TPWS was not designed to prevent a SPAD but rather to reduce the consequences of one.

 

[cjmillsnun]

ATO (Automatic Train Operation) is not a safety system, it is a means of letting a computer control a train's motors — usually by pressing a button when stopped in a platform, that causes the train to drive to the next platform and stop. It is only used on certain tube lines, and is unsuitable for national deployment.

The "best protection" is currently ATP (Automatic Train Protection), which uses a computer to continuously check that a train is being driven safely, which would allow it to be stopped prior to a SPAD. This was developed after the Southall and Ladbroke Grove crashes on the GWML in the late '90s. It was originally planned to be deployed nationwide but was cut back on cost grounds — but is currently used on the GWML and the Chiltern line.
There is a list of accidents that ATP would have prevented here.

It was fitted and available before either of those crashes. It was a BR trial on Chiltern and the GWML

 

[ac6000cw]

It was fitted and available before either of those crashes. It was a BR trial on Chiltern and the GWML

Quite correct.

It was recommended to be extended nationwide by the public enquiry that followed the 1988 Clapham accident (even though that was caused by a signalling wiring fault, so it would not have prevented that accident), but the politicians later baulked at the cost so it's never happened. A good example of the cost-versus-benefit situation.

Widespread, raoid, installation of TPWS instead was deemed to be a much better safety gain versus cost.

 

[edwin_m]

Widespread, raoid, installation of TPWS instead was deemed to be a much better safety gain versus cost.

TPWS is reckoned to have had about 60% of the benefits that ATP would have. It would have prevented Ladbroke Grove if it had been fitted, but was still awiting go-ahead from John Prescott at the time.

 

[ComUtoR]

You have got a safe railway though and the only reason I can see for you making this thread is because of the one about a train stopping short.

I have no doubts that the railway is safe and haven't said otherwise. Whilst the stop short thread did prompt my post my reasons are because I consistently see post after post, article after article, RAIB, major incidents, and various railways related issues and I think to myself 'Why aren't we preventing incidents ?' My post isnt a a concern over safety on the railway but one of a question why prevention isn't priority but reduction and mitigation is.

My personal thoughts about specific incidents are one that wouldn't be aired on a public forum

It strikes me you have a very simplistic view of the railway, it's needs and it's capabilities. It is the second safest bulk passenger transport system there is, the first being air travel.

It's not a simplistic view. It's one that comes to a forum where there is a very broad cross section of the industry and its onlookers. I like to know the reasons why and this forum is good at providing relevant insights.

As an aside to this, are you qualified to carry out detailed risk assessments? The railway has those very people and what you may perceive as a high risk is actually not that high in practice.

No and I have no intention to carry one out. I am however a tax payer and passenger. I also have a vested interest in the day to day running of the railway (personal reasons) I will also reiterate for those that have any doubts or misperceive my motivation for posting or my belief in the railway level of safety...

I have no doubts that its safe BUT are we making compromises that may bite us in the ass later down the line ? Is it really a matter of money and accepting that people may die (in extremis) or get injured so meh it happens but we saved a few quid ? If we have technology why aren't we using it ?

Well, of course they can. But you wouldn't be saying that if you used a 4-car service on a 12-car platform, with the exit at the rear of an uncovered platform

It's a pain in the bum but safety shouldn't be at the mercy of the lazy. If its reduction or removal of stop short incidents Vs my little legs or standing in the rain I know where my choice is. (I need to exercise anyway )


Cheers to all who have posted some interesting points have been raised.

 

[HSTEd]

There are safety implications of making people stand on freezing platforms in he rain as well.
And with modern equipment I would have hoped multiple train stopping points on each platforms would not be beyond he wit of man.

 

[Bushy]

A huge effort (and cost) goes into reducing risk on the railways to As Low as Reasonably Practical. Practicality includes consideration of cost and always will, as we do not, and never will, have infinite amounts of money to sink into railways or other infrastructure systems.
The aim of the RAIB

is To improve the safety of the railways by:

• Carrying out investigations to determine the causes and circumstances of railway accidents and incidents.
• Making evidence based safety recommendations, to reduce the likelihood and mitigate such causes and circumstances recurring in the future.

There is no industry where cost is not considered when assessing risk mitigation.
You can, of course, eliminate risk from railway operation - stop operating railways!

Regards

Bushy

 

[455driver]

That's the crux of it isn't it. How much and who pays. What about all those people who have lost their lives or children's lives because the attitude was cost.

I think a life is costed at £1million on the railway, IE if the improvement costs £999,999 and will save one life then the improvement will be done, if it costs £1,000,001 then it wont.
I think on the roads its £250,000!

Make of that what you will.

Is a Drivers safety of the line record based on % or signals/station stops ?

A drivers SOL record takes many many things into account (number of station stops isnt one of them ), some very small things can have a massive effect on it.

Instead of absolute safety on the railways maybe you should aim your ire at the roads where (if the figures are to be believed) another 4 people have died since I posted on this thread last night!
What about their families etc?
--- old post above --- --- new post below ---

I consistently see post after post, article after article, RAIB, major incidents, and various railways related issues and I think to myself 'Why aren't we preventing incidents ?

You see post after post because the incident is newsworthy!
Why is it newsworthy? Because it doesnt happen every day thats why!
Did you read in the papers about the 4 people that died (or the hundreds injured) on the roads yesterday? Or the day before? Or the day before? Or the day before? etc etc.
No? Why do you think that would be!

Edit-
If you want absolute safety on the railways then I will off down the job centre, because the only way you will ever get absolute safety is to close them down, no trains moving means nobody can get injured by them or on them!
Is that what you want, everyone on the roads where they are 500 times (or whatever the figure is) more likely to be injured or killed?
--- old post above --- --- new post below ---

There are safety implications of making people stand on freezing platforms in he rain as well.
And with modern equipment I would have hoped multiple train stopping points on each platforms would not be beyond he wit of man.

Assuming there are any stopping marks on the platform at all, I can think of several that dont have any marks and trains from 1 to 6 coaches will call at them!

 

[Moonshot]

What is there to stop the industry using axle counters on the track at the start of a platform which operate a platform only signal to indicate whether a train is fully accomadated on a platform or not? Not that I know the answer, but has this ever been thought of or tried?

 

[HSTEd]

The problem is the driver forgets how long the train is. Telling him how many coaches are on the platform wont neccesarily help much.

 

[Moonshot]

The problem is the driver forgets how long the train is. Telling him how many coaches are on the platform wont neccesarily help much.

yes ....that did occur to me after I initially posted......but I dont think I am a million miles away from a solution. So maybe just a platform signal only which indicates red or green operated in much the same way as automatic signals are on the track......Im sure one of the signal experts on here could probably suggest something....

 

[ommerson]

The only sane way to control risk is with normalised risk data. Such data could be fatalities per kilometre traveled by a mode of transport, or perhaps the annualised risk of death experienced by train drivers from level crossing incidents.

Engineers asses the efficacy of measures to mitigate risk with cost-benfit analysis. This is clearly a hugely emotive subject, and they don't like to talk about it because it would make for appalling PR. You just need to look to the National Institute for Clinical Excellent for evidence of this. NICE apply precisely the same kind of analysis to the effectiveness (of frequently, otherwise) of healthcare interventions.

The RAIB has been rather less careful in keeping the figures secret. See §86 of the Report into the Oxshot Incident . The limit for cost-effective interventions to prevent a railway fatality is quoted (for 2010) as £1.6m. Whilst it doesn't quality this, I presume this is over the asset life of the intervention. For public highways the figure is slightly lower at £1.5m.

By comparison, the figure for the civil aviation industry is likely much higher - not because of the liability for passenger fatalities - which are pegged really low by international convention - but rather for reputation reasons and the enormous cost of planes.

Meanwhile, the annualised risk of death for UK men aged 45-55 is about 1/279. Most of these deaths will be through illness (complications of diabetes, cancer, heath attacks, stroke) and motor accidents.
This is probably the mean age of many train drivers who are also exposed to additional risk factors relating to level crossing accidents. They also lead a sedentary occupation with disrupted sleep patterns from shift-work - a known risk factor for health.
Whilst I haven't got the normalised rate of death for train drivers in level-cross accidents data to hand, I bet the risk is considerably lower than 1:279 - probably by an order of magnitude or two.

Perhaps NR and TOCs would get a higher reduction in employee fatalities by investing in better occupational medicine services - particularly focusing on preventative medicine and perhaps providing free gyms and free healthy food in crew depots rather than in making level crossings safer?

 

[carriageline]

yes ....that did occur to me after I initially posted......but I dont think I am a million miles away from a solution. So maybe just a platform signal only which indicates red or green operated in much the same way as automatic signals are on the track......Im sure one of the signal experts on here could probably suggest something....

Yes but how often does this happen that you want to spend millions installing these to platforms? And a green or red signal wouldn't work for multitude of reasons, for example unless you completly add it to the interlocking, drivers would see irregular aspects everywhere. Maybe a "OD" (for open doors of course!) on the signal, the same as how CD/RA works. But even then, you still have room for error.

I think your trying to fix a problem that isn't really there. There are much bigger problems out there that could do with that money spending on it



Sent from my iPhone using Tapatalk

 

[SpacePhoenix]

Could the system they've got on the LU get adapted (the rectangular green boards), possibly with it telling the trains onboard computer how long the platform is that it's stopped at then the computer decides which doors to release?

 

[underbank]

Time and time again, not particular to the railways, we see unintended consequences
arising from an action which was a good idea with the best of intentions. We seem very poor at considering behavioural patterns resulting from what should be "good" decisions, whether it's H&S, raising/lowering taxes/benefits, etc etc. It's always the unintended consequences that bite them in the bum a little down the line.

OK, so replace a crossing with a footbridge - how would you feel if someone fell and killed themselves on the bridge steps? - That same person who'd been crossing the track safely for decades? OK, close down a road crossing. That turns a local housing estate into a rat run and some toddler ends up squashed under a car? A lot of the time, you're just changing the risk - yes, the "railways" would be safer, but overall, life around the railway is more dangerous in my two examples above.

As for level crossings, I am more and more convinced that the increase in idiots trying to jump over or drive under the barriers is an unintended consequence of trying to make them safer. It seems a trend for the barriers to be lowered sooner ahead of the train's approach, meaning that drivers and pedestrians know they're going to have to wait longer once they see the lights start to flash - some will not want to wait so will take the risk. Whereas in the past, the barriers/gates would close for shorter periods of time, meaning the road users were more likely to not risk it in the knowledge they'd only be waiting a short time. Then of course the scenario when someone is trying to catch the approaching train knowing that once the barriers are down, they can't cross the line to the other platform to get on it, so will try to get under the lowering barrier so they're on the right side - again, an unintended consequence of them being lowered potentially far earlier than they need to be.

Again re level crossings, it's simply impossible for some to be removed or bridges built instead. Many were historically located at platform ends in towns and villages, and whilst the stations may have been closed down, the site will have shops, offices, houses and other road junctions within just a few tens of feet, meaning that a bridge is simply impossible without major demolition of nearby properties, which just isn't going to happen. The alternative of closure can't happen either as some locations won't have an alternative reasonable route from one side to the other, meaning effectively killing off one side of the tracks.

 

[rf_ioliver]

'Stopping Short'
Why aren't trains fitted with gps and technology to prevent door openings when short of the stopping point ?
What about other tech ?
Why cant all trains stop at a single point ?

Because the addition of additional technology to solve that problem would bring its own risks. What happens when the GPS is inaccurate, when the system fails, what about all the related and interconnected systems?

'Level Crossings'
Remove them all and use bridges, tunnels, viaducts etc.

Again, each introduces its own set of risks. A tunnel might be a good idea but it comes with all the risks of a tunnel which themselves are mitigated by various technologies and procedures - which as you can guess introduce different risks

SPAD's
Can trains be stopped BEFORE the signal instead of within the overlap ?
Is ERTMS the solution ?
Whats the best train/signal protection (TPWS/ATP/ATO etc ?)

I though that SPAD signals were introduced to mitigate the SPAD risk by providing a second chance to stop with the blue signals with red flashing lights. ERTMS has its own complexities and similarly each protection mechanism is designed for a specific purpose. One of the reads AWS has been succesfull is that it is very simple, has a clear defined purpose and introduces minimal risk by offsetting issues with driver vigilence.

Indeed signalling systems are designed with this in mind. Consider the 4 aspect system. In the absence of a light assume signal at danger. If a yellow light fails then a double aspect will be shown as a single yellow aspect which is more restrictive. The AWS system is always triggered in a non-green scenario etc. It takes a lot of additional technology to really, effectively improve on this.

Overall, the balance is often not between cost and risk, but between risk and risk. Solving problems with technology always introduces further risk.

Take the aircraft industry as an example, most of the risks have been mitigated not by technology but by human procedures. Indeed most training on safety focusses on the role of humans in the loop (CRM) and taking over in situations where the technology fails. There is a very good video called 'Children of the Magenta' which is a training presentation on safety to pilots made by American Airlines; there used to be a few copies on YouTube before AA pulled them - but well worth watching if you can find it.

I've worked with safety critical systems and even update, change, 'improvement' or otherwise needs to be checked and cross-referenced with all kinds of interesting analyses: RCA, FMEA etc. My job at the moment is applying these in systems that aren't usually considered 'safety-critical' ... now that's fun!

The railway industry, just like medicine, aviation etc is extremely advanced in designing systems for safety and very, very rarely does cost factor in a decision of this nature until very late in the process. Most of the time the unintended consequences will counterbalance any introduction of further safety systems.

Consider manual opening of doors (how did we survive in the 1980's and earlier?), which is better?

Place a notice on the door stating "Do not open while train is in motion. Wait until train has stopped at the station before opening. Do not lean on the door"

versus

Automatic opening based upon locks that are released according to a GPS location.

In the first, you just know that someone is going to open the door and fall out - whose fault is that...it is very hard to protect from human stupidity. In the second case consider all the systems that the door locking mechanism need to be interact with. How do you handle failure of any system in this case. What happens if the GPS system is inaccurate or fails? What happens if the guard can or can not override the system? What happens in an accident or power failure etc?

t.

Ian

 

[carriageline]

Fairly good post Ian, although to be a pedant, a double yellow with one of the yellows failing won't necessarily show a single yellow, (infact I doubt any will due to incorrect aspect sequence) will just show a blank.

And with the advent of TPWS, I believe SPAD signals are being removed, I don't think there is many (if any at all?!) been added in recent years.


Sent from my iPhone using Tapatalk

 

[ac6000cw]

I have no doubts that its safe BUT are we making compromises that may bite us in the ass later down the line ? Is it really a matter of money and accepting that people may die (in extremis) or get injured so meh it happens but we saved a few quid ? If we have technology why aren't we using it ?

I don't know what sort of job you do, but my background is doing engineering design, and one thing about designing products, systems etc. is that it is always a compromise, taking into account all sorts of factors - money is always big factor (no one has an unlimited budget, or time, or facilities to build and maintain a product). We have to operate in the real world of limited resources and a lot of competition for them, and we all die eventually - it's just a matter of when.....

 

[Moonshot]

Yes but how often does this happen that you want to spend millions installing these to platforms? And a green or red signal wouldn't work for multitude of reasons, for example unless you completly add it to the interlocking, drivers would see irregular aspects everywhere. Maybe a "OD" (for open doors of course!) on the signal, the same as how CD/RA works. But even then, you still have room for error.

I think your trying to fix a problem that isn't really there. There are much bigger problems out there that could do with that money spending on it


Sent from my iPhone using Tapatalk

judging by the response on the Northern/TPE consultation thread, every other poster seems to think its an issue!!!

Im sure there is a simple tech solution to what is percieved as an issue of doors released by driver when not all accomadated on platform. As others have said though, the cost may well outweigh the benefit......lets be honest here, just how many passnegers have fell off trains and become a fatatlity?

 

[ac6000cw]

As for level crossings, I am more and more convinced that the increase in idiots trying to jump over or drive under the barriers is an unintended consequence of trying to make them safer. It seems a trend for the barriers to be lowered sooner ahead of the train's approach, meaning that drivers and pedestrians know they're going to have to wait longer once they see the lights start to flash - some will not want to wait so will take the risk. Whereas in the past, the barriers/gates would close for shorter periods of time, meaning the road users were more likely to not risk it in the knowledge they'd only be waiting a short time

That's certainly a major factor in level crossing misuse in the US and Canada, where freight trains (the dominant traffic in most cases) are much longer, so if a train is going upgrade at 10-20 mph it can mean waiting for 5 minutes or more for it to go past. Given that over there crossings are almost always AHB (or 'open'), it's an easy temptation to weave through the barriers....

 

[Clip]

I have no doubts that the railway is safe and haven't said otherwise.

Oh? I thought you did when you said this

I have to take the attitude that if I want a safe railway and I'm travelling on a train then I am willing to pay for it. My son is at an age where he will be travelling on a train by himself. I want that train safe at whatever the cost. ?

That to me reads as you are willing to pay more for a safe railway indicating that it is not safe already.

As someone has pointed out - you read threads of incidents on here because they are an issue(no speculating though) on an otherwise extremely safe mode of transport.

 

[fowler9]

Oh? I thought you did when you said this



That to me reads as you are willing to pay more for a safe railway indicating that it is not safe already.

As someone has pointed out - you read threads of incidents on here because they are an issue(no speculating though) on an otherwise extremely safe mode of transport.

Yeah to be fair the railways in this country are already an incredibly safe way to travel. I'd be more worried about a child of mine reaching an age when they can get in to a car driven by a mate or even themselves.

 

[455driver]

Anyone remember the thread about 'my idea to remove 3rd rail as a priority' which then didnt have any actual details about how it would be done!
The bit in that thread about the number of people killed by the 3rd rail (on a daily basis it seems) was actually quite funny in its lack of detail!

 

[ac6000cw]

Yeah to be fair the railways in this country are already an incredibly safe way to travel. I'd be more worried about a child of mine reaching an age when they can get in to a car driven by a mate or even themselves.

Quite... it's why insurance is so expensive for young drivers.

 

[rf_ioliver]

Fairly good post Ian, although to be a pedant, a double yellow with one of the yellows failing won't necessarily show a single yellow, (infact I doubt any will due to incorrect aspect sequence) will just show a blank.

And with the advent of TPWS, I believe SPAD signals are being removed, I don't think there is many (if any at all?!) been added in recent years.

Thanks. I'm not familiar on the actual implementation UK signalling, however in this situation you've actually got to fail-safe modes, both where a failure causes a more restrictive aspect to be show, ie: single yellow or nothing; so you end up even covering the situation where an incorrect sequence detection fails. Without knowing specifics I'm guessing it was designed that way.

Actually I've met some of the people working on formal specification of parts of the French signalling systems and had the pleasure of talking with one of the designers of the signalling system used on the French LGV lines - that was fun!

t.

Ian

 

[Railsigns]

Fairly good post Ian, although to be a pedant, a double yellow with one of the yellows failing won't necessarily show a single yellow, (infact [sic] I doubt any will due to incorrect aspect sequence) will just show a blank.

If a signal was showing double yellow and one of the yellows fails, it will show a single yellow.

 

[edwin_m]

If a signal was showing double yellow and one of the yellows fails, it will show a single yellow.

The lamp proving circuit will also replace the previous signal to danger in this situation.

Yet another safety protection is provided by signal lamps having double filaments. If the first filament fails the system changes over to the second one and also sends an alarm to the technician. So it would be very unlikely for the lamp to fail totally before it was replaced. LEDs are a bit different but have equivalent detection circuits and of course last a lot longer as well.

 

[Railsigns]

The lamp proving circuit will also replace the previous signal to danger in this situation.

Failure of the top yellow (the one not used to display a legitimate single yellow aspect) doesn't affect the previous signal.