Safety Culture, in the UK and the USA and railways/railroads

[Glorious NSE]

Should the warnings/preventative measures be initiated with enough time for road users (including pedestrians) to get out of the way, or with enough time for trains to stop?

 

If you did it to train stopping time then you drastically increase the chance that someone will get fed up waiting and go round the barriers.

[Edwin_m]

If you did it to train stopping time then you drastically increase the chance that someone will get fed up waiting and go round the barriers.

And you need some way of detecting the crossing to be clear, otherwise there is no way of knowing whether to stop the train or not.  This is exactly what Network Rail is now implementing with their obstacle detection systems.  They should be safer but one consequence will be much longer road closure times so as to give the train a clear run on green signals.  These crossings do at least have full barriers to prevent zigzagging. 

[dave1905]

A question about level (grade) crossings.

 

Should the warnings/preventative measures be initiated with enough time for road users (including pedestrians) to get out of the way, or with enough time for trains to stop?

 

Put another way, how much responsibility for people's actions should the whole system take away from them? How much should be left to people using their nous?

 

The train stopping is irrelevant.  The trains don't stop at the crossings, people and vehicles do.  The train takes a mile or more to stop.  At 30 mph it takes 2 minutes to travel the mile  to stop the train.  That means you would have to know two minutes before the train gets there that the crossing will be blocked.  There isn't any system that can tell you that the crossing will be blocked two minutes from now.  The current warning times are enough, the problems occur when somebody tries to beat the train across the tracks (in which case no amount of warning time is enough), somebody pulls on a crossing and fails to get off in time (better coordination with traffic lights would solve some of that, better road construction the rest, both of those are a government function) or gets stuck on the crossing (wrong turn, high centered, etc) in which case no amount of warning time is sufficient (at that point the road traffic should use the emergency numbers at every crossing to notify the railroad).

[Edwin_m]

Some facts that will help (courtesy of the CIA World Factbook) - the length of the railways in the respective countries:

 

UK (excluding NI) - 16,151 km

 

Canada - 46,552 km

US - 224,792 km

 

So on a simple basis of you can expect x accidents per 1,000 km per year, the size of the system in North America means in absolute numbers it is going to seem much worse than it is in the UK.  So for a hypothetical 1 accident per 1,000 km that means while the UK would end up with 16 accidents, North America would have 271 accidents which would make it "seem" more dangerous in North America.

 

It also demonstrates why some of the options in the UK - like fencing the railway in - really aren't feasible over here.

 

As for level crossings, again the differences are significant.  Wikipedia states there are 1,600 road crossings in the UK, while according to (1) there are 40,000 in Canada - I can't quickly find any numbers for the US - and the average cost of grade separating a crossing is $40 million.

 

Just checked Google Earth, and here in Mississauga there are 19 level crossings on the 2 major rail lines - or $760 million to grade separate them all.  Some could instead be eliminated, but your still talking at least $500 million in one city.  Where is that money going to come from?

 

And of course, if you could come up with that money would there be better ways to spend it.

 

Canada averages 216 grade crossing accidents per year, with 27 fatalities.  In many ways an unacceptable number, but there are 2,077 deaths on Canadian roads each year(2).  In that context, the grade crossing isn't an issue.

 

I am not saying that there is nothing that can be done, or should be done.  Certainly the crude oil seems at the moment to be a particular issue that needs to be examined, but the number of crossings and the amount of track in North America means you are never going to make the system entirely safe.

 

 

(1) - http://www.cbc.ca/news/canada/ottawa-bus-crash-how-safe-are-train-level-crossings-1.1861017

 

(2) - https://www.tc.gc.ca/media/documents/roadsafety/cmvtcs2012_eng.pdf

It's very difficult to make a fair comparison of crossing safety because of several factors including:

• Frequency of trains - on most crossings in North America there are probably fewer (but bigger) trains than on most in the UK
• Frequency of road vehicles - many crossings in deep rural areas in NA will have far less than the UK average as a result of the lower population density.  However those in NA urban areas could have far more, as my perception is that crossings of busy multi-lane highways are uncommon in the UK but quite common in NA

The product of those two, summed over all crossings, could be compared with the actual collision rates to give an idea of the probability of collision, normalised against the number of times a road vehicle potentially encounters a train.  Even this isn't perfect, as a crossing with fewer trains may actually be less safe because road users may assume there is no chance of a train arriving.  A we don't know any of the number such a comparison isn't possible, but it gives an idea of how difficult a meaningful comparison is.  RSSB has done at least one comparison of crossing safety records across Europe, where conditions are more similar and incidentally the UK comes out as one of the safest.  But I'm not aware of anyone doing a statistically valid comparison across the Atlantic. 

 

Published statistics do suggest that full-barrier crossings where the track is proved clear of obstructions before clearing the signals are very much safer than the automatic types that don't do this, and that any sort of barrier improves the safety record significantly over just lights.  Since the maximum protection at North American crossings appears to be approximately the same as our AHBs then, other things being equal which of course they aren't, the safety record is likely to be worse in NA. 

 

Interesting from the above figures that the total road fatalities in Canada are similar to those in the UK.  Canada has roughly half the population but far more land area, and I would expect people to drive

[Edwin_m]

In terms of safety culture I suggest level crossing accidents need to be taken out of the equation.  Certainly Amtrak has little influence over equipment fitted or driver behaviour at most of the crossings it uses, for the simple reason that they are owned by the freight railroads.  And across North America custom and practice at crossings may (discuss?) depend more on the attitude of the local authorities than that of the railroad.  Picking up on something in the back of my mind when writing my two previous posts, is there any interest in NA in obstacle detection systems, coupled with signal protection, as now used by Network Rail?  I know NA has used crossing predictor systems for some time to adjust the light/barrier sequence to the speed of the approaching train, which suggests they are sensitive to closure times either as a safety issue or to minimise highway disruptions.  There was also a move towards four-quadrant barriers (full barriers in UK speak) a few years ago, but I think still operating automatically rather than with proving of the crossing clear as in the UK. 

 

On the wider attitude to safety I wonder if there is a distinction between the Class Is and the various short lines.  Again only perception but the latter seem to be operated on a shoestring with little regard for safety, the Lac Megantic disaster being the most memorable result.  But are the big operators more safety-conscious?  Perhaps they need to be as they are large and wealthy corporations and tempting targets for lawyers, and would probably be sued out of existence if found to be blameworthy for a few major incidents.  A short line in the same situation would just go to the wall and perhaps be reborn with a new identity.  I believe this is an important distinction across the Atlantic, that Europe relies on health and safety legislation but to a large extent the same result is achieved in North America by the threat of legal action if something goes wrong. 

[Adrian Wintle]

On the wider attitude to safety I wonder if there is a distinction between the Class Is and the various short lines.  Again only perception but the latter seem to be operated on a shoestring with little regard for safety, the Lac Megantic disaster being the most memorable result.  But are the big operators more safety-conscious?  Perhaps they need to be as they are large and wealthy corporations and tempting targets for lawyers, and would probably be sued out of existence if found to be blameworthy for a few major incidents.  A short line in the same situation would just go to the wall and perhaps be reborn with a new identity.  I believe this is an important distinction across the Atlantic, that Europe relies on health and safety legislation but to a large extent the same result is achieved in North America by the threat of legal action if something goes wrong. 

 

In terms of level crossings, the distinction is basically moot. As long as the crossing is signed and any signals/barriers are working, the train company is not at fault, so there isn't really a threat of legal action (there are exceptions, but the railroad would likely be sued along with the other stakeholders in the crossing - state, municipality, etc). Basically, being on the tracks at that point is trespass.

 

Note that, except in certain urban areas, the train is also whistling as it approaches the crossing, and it is hard to miss that.

 

Adrian

[Ozexpatriate]

Note that, except in certain urban areas, the train is also whistling as it approaches the crossing, and it is hard to miss that.

Yet people do, and the flashing lights as well.

[Ozexpatriate]

On the wider attitude to safety I wonder if there is a distinction between the Class Is and the various short lines.  Again only perception but the latter seem to be operated on a shoestring with little regard for safety, the Lac Megantic disaster being the most memorable result.  But are the big operators more safety-conscious?

First, not all short lines are the same. In general, railroad operators (both Class I and on short lines) are safety conscious and there are regulations - both legislative and operational for safe operation.

 

The Lac Megantic tragedy opens the scope to many questions including but not limited to oil tanker structural integrity, distance and the integrity of the permanent way, locomotive maintenance, the practice of parking a train on a grade on the mainline, the rules for tying down a train on a grade, and many other factors which are relevant to the general safety topic but not limited to one thing.

[Ozexpatriate]

The NTSB has a 'most wanted list' that includes:

• Disconnect from deadly distractions (focus is drivers)
• End substance impairment in transportation
• Enhance public helicopter safety
• Implement positive train control in 2015
• Improve rail tank car safety
• Make mass transit safer (light rail, commuter rail, subways, ferries, streetcars,

  buses and trolley buses)

• Prevent loss of control in general aviation
• Require medical fitness for duty
• Strengthen commercial trucking safety
• Strengthen procedural compliance (aviation)

Note that the only one with a date is positive train control. The Rail Safety Improvement Act of 2008 mandates implementation of a PTC system by each Class 1 rail carrier and each provider of regularly-scheduled intercity or commuter rail passenger service by December 31, 2015.

[Simond]

Interesting debate.

 

Couple of points from my localities.

 

Romney Hythe & Dymchurch has lost two drivers to crossing accidents that were, in my opinion, entirely due to road users failing to obey valid traffic signals. I wasn't in court, I didn't hear the evidence, but I remaining astonished that neither were punished for manslaughter.

 

I lived in india for three years. On average, 10 people died on the railways around Mumbai every day because they couldn't be bothered to use authorised crossing points/footbridges, but ran across the tracks, presumably in front of oncoming trains. Darwin rules...

 

Doesn't help the poor sod in the cab though.

 

And given what happened in the Alps today, you can be unlucky too.

 

Best, live long, be safe.

Simon

[Talltim]

I think you miss my point, but that may (as I stated earlier) be because you are in the USA, and I am speaking from the UK, where such things can and do happe, and it is impossible for someone to have in-depth contemporaneous knowledge of both countries. My local level crossing has seen a big increase in the amount of time between the gates (well, barriers, of which there are four) coming down and trains crossing, precisely for this reason - too many pedestrians taking their life into their own hands whilst endangering others - and also because there were a couple of "run through" incidents where trains were supposed to stop because the barriers were up, but didn't.

The point being that there are now crossings in the UK where the train stops for the crossing, but only only in situations where the equipment has failed (lights or barriers) or where the crossing cannot be detected to be clear. Radar and LIDAR is used for this. CCTV is also used in some situations where the signaller visually checks the crossing before giving the train the green 

[jjb1970]

On the particular subject of human error I qualified this by referencing industrial accident reports. I've investigated quite a few incidents and investigated a lot more and common recurring themes that I see when human error is used as an explanation include:

 

An over reliance on procedural controls in known high hazard activities, it is pretty much universally acknowledged that procedural controls should not be the principal means of risk control in a high hazard process. Where procedures are necessary (eg. maintenance, plant start up and shut down), are they usable and fit for their intended purpose? I've read procedures that were absurdly long and complex and an invitation to be ignored and plenty of others that were clearly written by people who'd never tried to do the jobs covered.

 

Inadequate training, it is easy to dismiss somebody who does something that strikes us as silly as some sort of idiot but what training have they received? Did they receive any training? How were they assessed as competent to do the job?

 

Work place culture, a pro-active "can do" culture is great but there is a line between a positive approach like this and stepping over into the realms of recklessness. Managers should have a very good idea of how long a job should take, if they see that a night shift has done a job remarkably quickly do they go and ask how it was done so they can pass learning points on to everybody else, do they have a quiet word and say they appreciate people want to get things done but it won't happen again or do they give a wink and a nudge and say "good job guys"?

 

Lack of resource leading to excessive workload and stress leading to mistakes or all but forcing people to take short cuts and develop workarounds. Or staff who have been recognised as suffering from stress, fatigue or medical conditions but allowed to keep going.

 

Poor ergonomics, I've seen control system layouts that were an open invitation to error. One of my favourites was one ship where the diesel generator engine stop/start and breaker open/close buttons were offset so that the breaker controls for generator 1 were directly below the engine stop/start for engine 2. I think every engineer on that ship caused a black out by stopping the wrong engine once by just stopping the engine above the breaker switch, an entirely predictable consequence of very bad design.

 

Poor asset condition leading to all sorts of weird and wonderful workarounds just to try and make stuff work.

 

Lack of plant information, drawings which are not as per the state of the plant (eg. P&ID's not kept up to date), information not available in the working language of the crew (a big problem on ships) and no common language to facilitate communication.

 

Cost cutting that does not provide the right tools and equipment to do jobs safely

 

I've seen so many incidents that put human error in big letters front and centre and even the most cursory review will reveal that it was not human error but more fundamental problems. And a fundamental point  to remember is that in most cases the victims in these cases that are held to be human error are not idiots, not stupid nor reckless per se but normal workers who do something that probably most people in a job have done but have gotten away with but who draw the short straw and pay a heavy price. We all know that people will fail at times for a whole variety of reasons, people make mistakes, yes people sometimes do dumb things and safety systems should be designed to take that into account and designed to try and protect people.

[Gerald Henriksen]

Picking up on something in the back of my mind when writing my two previous posts, is there any interest in NA in obstacle detection systems, coupled with signal protection, as now used by Network Rail?

As Dave said, the size of trains in North America is such that you need 2 minutes or so to stop a train. You can't know 2 minutes in advance whether the crossing is blocked or not (in most cases), so obstacle detection systems are of no use.

 

On the wider attitude to safety I wonder if there is a distinction between the Class Is and the various short lines.  Again only perception but the latter seem to be operated on a shoestring with little regard for safety, the Lac Megantic disaster being the most memorable result.

I think it is dangerous to equate the operations of all shortlines with how one particular shortline operated.

[Ozexpatriate]

I lived in india for three years. On average, 10 people died on the railways around Mumbai every day because they couldn't be bothered to use authorised crossing points/footbridges, but ran across the tracks, presumably in front of oncoming trains.

Your reference to trespass casualties in India connected me in part to why I included the "Human lives are cheap" observation that Sir Felix Pole highlighted.

 

I'll make the observation that most humans respond to tragedy in ways that are inversely proportional to distance and cultural separation. A single fatality in our own community touches us much more closely than multiple fatalities far away. A ferry overturning in Indonesia might not necessarily feel close to home nor might really large scale disasters like a flood in Bangladesh that kills thousands. With all the dreadful things that happen in the world a uniform level of empathy for each of them would be debilitating.

 

I do think that the magnitudes of population and distances in the US are relevant to this kind of self-preserving detachment and it is relative.

 

I'm struggling with the wording, but I think an aspect of safety culture can mitigate some of that detachment.

[dave1905]

In the US modern crossings have circuits that detect if the train stops and put the gates back up. If they train starts up again the system detects that and puts the gates back down.

 

If the gates are damaged or malfunction the trains are given appropriate instructions which could entail stopping and flagging the crossing.

[Talltim]

As Dave said, the size of trains in North America is such that you need 2 minutes or so to stop a train. You can't know 2 minutes in advance whether the crossing is blocked or not (in most cases), so obstacle detection systems are of no use.

I'm not sure size, or weight, of trains is the only factor in stopping distances. Surely it has more to do with speed and brake force. After all, a light engine has a lower speed limit than a ten coach train, because the brake force* is lower and it can't stop as well. A twenty car train should stop as well as a 200 car train.

There are other factors with air brakes, such as the speed of application, and here, yes, train length makes a difference. If an EOT device that can release the air pressure is used, it halves the time for all the brakes to apply.

*I'm never sure if brake force is purely a measure of the power of the brakes, or whether it is a ratio of power of the brake to weight (or mass if you want)

[rockershovel]

I'm not going to comment on specific safety procedures in US railroads, not my field; but I will say, from experience if the oil industry, that American practices are much more likely than British practices to result in preventable incidents, because the pressure to take commercially (cost) driven decisions is much higher.

 

Look at the record of Amoco (now BP) in the US, from Macondo to Texas City and Alaska, compare it to BP elsewhere and draw your own conclusions

 

I certainly saw an example if this in British railway practice; the influx of insufficiently trained, often immigrant labour onto the railways, sometimes lacking sufficient language skills, following privatisation and tbe attendant destruction if the established track msintenance structure.

 

This is something you see a great deal of in US practices, because its a US thing; 5% of Americans move to another state seeking or following work annually, and there is no integrated national training scheme in any industry. Working with limited labour availability, high turnover and big gaps in the skills base is a defining feature of US practice and has never been any different.

 

Same thing goes for Indonesian ferry accidents and floods in Bangla Desh. I know, from experience, that Indonesian ferries are often very dangerous compared to European ones because of fundamental differences, one if whuch us that Asuan cukture does not place the same value on human life. I know, from experience, that much of Bangla Desh consists of river deltas which flood regularly. I know, from experience, that these issues are rooted in uncontrollable popukation growth and grinding poverty in which all resources are immediately consumed and however many die, population growth remains uncontrollable.

 

It isn't simply distance, it's the feeling that given the context and environment, these thungs are inevitable. Asian fatalism isn't a sign if cuktural degeneracy; it's the natural outcome of living in societies which must cope with pressures of this sort, because they have no option.

[Glorious NSE]

the problems occur when somebody tries to beat the train across the tracks (in which case no amount of warning time is enough), somebody pulls on a crossing and fails to get off in time (better coordination with traffic lights would solve some of that, better road construction the rest, both of those are a government function) or gets stuck on the crossing (wrong turn, high centered, etc) in which case no amount of warning time is sufficient (at that point the road traffic should use the emergency numbers at every crossing to notify the railroad).

Sorry Dave, thats exactly my take on the US railroad view of crossings, it says in short, 'this is not our problem'.

 

And so we get regular repeats of the exact same utterly preventable wreck like the one I mentioned earlier. Not our problem.

 

The bit that frustrates me most is it leaves their train crew in the front line for this, I'd have expected more respect for them.

[Adrian Wintle]

I'm not sure size, or weight, of trains is the only factor in stopping distances. Surely it has more to do with speed and brake force. After all, a light engine has a lower speed limit than a ten coach train, because the brake force* is lower and it can't stop as well. A twenty car train should stop as well as a 200 car train.

There are other factors with air brakes, such as the speed of application, and here, yes, train length makes a difference. If an EOT device that can release the air pressure is used, it halves the time for all the brakes to apply.

*I'm never sure if brake force is purely a measure of the power of the brakes, or whether it is a ratio of power of the brake to weight (or mass if you want)

 

Ultimately it comes down to the coefficient of friction between steel (wheel) and steel (rail). Yes, you can improve things up to the point that you stop the rotation of the wheels and steel starts sliding on steel, but beyond that point you can't really improve things.

 

Adrian

[Adrian Wintle]

Yet people do, and the flashing lights as well.

 

They drive through down barriers and they drive or walk into moving trains as well. Ultimately there is a limit to how much protection can be put in place to protect people from themselves.

 

Adrian

[Grovenor]

 

and even the most cursory review will reveal that it was not human error but more fundamental problems.

You are using "human error" in a very restricted sense here are you not. Everyone of the other fundamental problems you cited are also human error, or are you assuming that the procedure writing, system designing, training, maintenance standards etc are all produced by non-humans.

Keith

[dave1905]

You are correct, there are a lot of things that are "not our problem".  If you see a neighbor on a rickety ladder using a frayed extension cord to do a home improvement, you could go into his house, confiscate the ladder, confiscate the cord and replace them.  Unfortunately  that would be trespass and theft. That would be against the law.  You as a private citizen have no authority to compel him to use a new ladder and cord. 

 

The railroad doesn't own the road.  We have no authority or responsibility to rebuild the road structure.  We have no legal authority or right confiscate the property on the other side of the road to move the road a safe distance from the tracks.  We have no legal right to evict the people and businesses living on the property.  We have no right or authority to compel the people living in the area to pay for the improvements or the maintenance of them.  We have not right or authority to install traffic control devices.  We have no right or authority to interlock anything on our right of way with the signal system on the public road.  The railroad as a corporate citizen has no right to compel the government or the adjacent property owners to improve their property and we have no legal authority to do it ourselves.

 

So if that to you is "not our problem" then yes, its not our problem.  Surprise!!! The ability to affect solutions has limits.  If we can't impose a solution then yes, its not our problem.  We can advise the owners of the problem on possible solutions (we do), we can take what steps we have authority to do (we do), but at some point there is a limit to what we can do.  After that, yes, it is not our problem.  I'm sorry reality has surprised you.

[jjb1970]

You are using "human error" in a very restricted sense here are you not. Everyone of the other fundamental problems you cited are also human error, or are you assuming that the procedure writing, system designing, training, maintenance standards etc are all produced by non-humans.

Keith

 

Indeed, a totally valid point, however they are not the humans who get fingered in incident reports that ends up being loaded up with the blame (noting that they generally include a disclaimer that the body producing the report believes in a no blame culture). This point is equally valid to efforts to reduce human failures by introducing fully automated processes, it moves the potential from errors from operators to those developing the software etc but the potential for human failure is still there.

[Reorte]

So if that to you is "not our problem" then yes, its not our problem.  Surprise!!! The ability to affect solutions has limits.  If we can't impose a solution then yes, its not our problem.  We can advise the owners of the problem on possible solutions (we do), we can take what steps we have authority to do (we do), but at some point there is a limit to what we can do.  After that, yes, it is not our problem.  I'm sorry reality has surprised you.

The problem with "not our problem" is that it carries implications of "and therefore couldn't care less", which I'm sure isn't the case.

[Glorious NSE]

You are correct, there are a lot of things that are "not our problem".  If you see a neighbor on a rickety ladder using a frayed extension cord to do a home improvement, you could go into his house, confiscate the ladder, confiscate the cord and replace them.  Unfortunately  that would be trespass and theft. That would be against the law.  You as a private citizen have no authority to compel him to use a new ladder and cord. 

 

No, that is him doing something as a private citizen which most likely does not affect anyone else. That is not the situation with a grade crossing.

 

If your dodgy neighbor was using his rickety ladder to cut down a tree that was adjacent to your property in a way that you felt compromised your personal safety, would you sit there and wait for it to crush your house, as obviously it wouldn't be your problem?