On Sat, 10 Mar 2007 13:46:14 +0000, JNugent wrote
(in message
pipex.net>):
>>>> Which is largely because of the rarity (and also because deaths in air
>>>> transport tend to occur in larger numbers in any one incident). In the
>>>> UK
>>>> alone I beleive we have something like as many deaths in 3 months on our
>>>> roads as there are worldwide by air - yet because they are an 'everyday
>>>> item' AND generally happen in small incidents then they seldom get
>>>> anything
>>>> more than a 60 second slot in the news.
>>>> I think that this disparity in reporting, which also afflicts trains, is
>>>> in part for the public clamour to DEMAND that rail and air travel be
>>>> made
>>>> 'safe' - as if anything can be made 100%% safe.
>
>>> Hang on...
>>> Much is made of the closed system that it is the
>>> National Train Set.
>
>> And that is relevant to the thread in what way ?
>
> Eh?
I made a comment about the way accidents are reported, you responded with a
comment about it being a closed system - how are the two related ? What does
the latter have to do with the comment about the way train crashes are made
into big news events because of their rarity ?
>>> It must surely take a very special sort of
>>> incompetence for a train to collide with anything
>>> else.
>
>> No,
>
> Stop.
>
> You are claiming that railway rashes can occur
> even when every person connected with the crash
> has behaved competently?
Yes
> Are you sure?
Yes
> No-one would claim that for a road crash.
>
> [BTW: I exempt both systems from charges of
> incompetence where the crash is caused by
> mechanical failure - other than where that happens
> because of the lack of routine required maintenance.]
Ahh, so by "crash" you don't mean "crash" !
I mean "crash" in the sense of "train comes off rails or whatever and hits
the front pages and primetime TV news", you mean "crash" as something else !
>> the fact that it is a closed system makes it much easier to reduce many
>> of the risks - but there is no such thing as risk free. Anyone who claims
>> that something can be made "risk free" is a complete moron !
>
> It can't be made risk free. A passenger may
> stumble and fall under a carriage. But crashes
> between trains are obviously caused by
> incompetence or negligence (if nothing else,
> negligence in not designing the signalling regime
> to fail to safety).
You contradict yourself there. You admit that things cannot be made risk
free, bujt then go on to say that it's incompetence to design a signalling
system that is not risk free. Make your mind up !
>> When you look at rail incidents, we have a mix of causes :
>
>> We had the infamous 'invasion of the system' where someone parked their
>> Land
>> Rover on the track. Apart from level crossings, this is a rare event and
>> is
>> generally not an easy thing to do - in this case the guy managed to miss
>> the
>> end of the safety barrier and have enough momentum to get all the way to
>> the
>> track
>
> ...where it was was - eventually - hit by a train
> travelling too fast to stop in an emergency - a
> systemic failure based on a fallacious premise.
> But we'll let that pass.
Why ? The whole point of having the closed system with it's signalling is
specifically to allow trains to operate at useful speeds instead of the 'stop
in distance seen to be clear' speeds they would otherwise have to use. Given
the limits of steel-steel grip, especially in the presence of contaminants
like ... err ... that rare thing in this country ... rain, then such speeds
would be so low as to make the railways virtually useless.
In general the concept of running a closed system works reasonably well, and
apart from AHB level crossings, incidents caused by incursion into the system
are fairly rare. I personally think the balance isn't far off - try and
reduce the risk further and I beleive the cost will escalate significantly
for limited benefit.
As for the problems caused by clueless f***wits abusiing AHB crossings, well
I think that's a different matter.
>
>> Mechanical failure. Hatfield was the most known and was caused by a broken
>> rail, but wheels can break, bearings fail, and various other things. You
>> can
>> reduce, but not eliminate, these by a good maintenance regime. You CANNOT
>> eliminate them altogether - after all, it's perfectly possible for a brand
>> new rail to have an embedded defect that causes it to fail a short time
>> after
>> laying. If you wanted to X-ray all the rails then the cost would be
>> phenominal and you would certainly reduce the accident rate - there'd be
>> no
>> trains running because no-one would be able to afford to use them !
>
> That's good perspective. I agree that mechanical
> failure is always a possibility. It's funny the
> way the usual suspects were baying for the blood
> of the track maintenance company's directors,
> though - isn't it?
Again, it's a symptom of the way these things are reported. Hatfield was THE
news for ages, in fact probaly for longer than the time taken to match the
death and injury toll by accidents on the roads. Because of this, and the
apparent inability of the 'general public' to consider risk in a rational
manner, it gets blown up into a "this can NEVER EVER UNDER ANY CIRCUMSTANCES
NO MATTER WHAT THE COST happen again" campaign - and then the customers
comlain when the fares go up to pay for it !
Perhaps we compulsory 'risk' appreciation lessons at school. IIRC, outside of
science and perhaps maths, risk and/or probability is not covered (or wasn't
when I was at school) and so non-science people really never get taught
anything about risk, risk factors, etc - they just go out into the world with
a belief that "risk is bad" and even worse, a belief that it's possible to
make something "safe".
>> Related to mechanical failure is 'system failure' - mostly signalling.
>> Yes,
>> the signalling system is all interlocked, but it relies on a large number
>> of
>> components all working correctly - including human elements. Yes the
>> systems
>> are designed to account for the human element, but unfortunately failures
>> do
>> occur.
>
> Surely the system should be designed to fail to
> safety?
To a point yes. And to a point it is. You still have the law of diminishing
returns so the more risk you want to eliminate, the more it costs. Also, bear
in mind that if you introduce additional complexity into the system so as to
create this 'fail safe' over a wider range of situations, then you are
introducing more points of failure.
> But you avoided mentioning Clapham and Paddington.
> Or Hither Green. Or Lewisham. Well, not
> specifically - you just skated over the worst UK
> rail accidents in living memory by obliquely
> referring to "'system failure' - mostly signalling".
Well I'm a northerner, so (for example) the one last week and the Shap one a
couple of years back are 'local' to me. Some of the ones you mention I can
remember now you've mentioned them, some I don't recall.
> IIRC (and I think I do) one (maybe more than one)
> of those terrible incidents occurred because a
> driver simply drove through a stop signal? How
> could that happen?
Failure of the optical link from signalling system to train controller which
therefore failed to stop the train - or, the driver didn't see the signal ! A
failure of the human element.
Incidentally, I think it would be useful if the railway people (whoever they
are these days !) did a video so that 'the man in the street' could see what
is involved from the drivers point of view. From shots I've seen on telly,
some of these signalling gantries are what could be described as 'busy' -
multiple signals for the multiple tracks, some of them on curves which makes
it less obvious which signal is which, and at complex junctions there are
multiple closely spaced gantries. All this adds up to a lot of signals and
scope for confusion - possibly an area ripe for reconsideration.
> And yes, before you say it, I also recall that
> here (and elsewhere), there were those who blandly
> assured us that driving trains through stop
> signals was routine.
>
> The point is that the worst accidents occurred on
> a closed system, with no-one to blame except
> railway workers and railway maintenance workers.
So, you want a totally risk free system - that seems to be your argument.
I'll guarantee that you aren't prepared to pay for it !
The system DOES evolve to reduce risks (such as those from drivers failing to
see a signal). I believe on at least the inter city routes they have an
automatic system( ATP, automatic train protection) that will stop the train
if it passes a red signal without the driver explicitely overriding it. I
don't know how far down this goes, and it may have limited applicability on
the very complex tracks of the commuter lines.
AIUI, one of the problems with ATP is that it applied not just to stop
signals, but also to caution signals. I imagine that when it was designed, it
was thought a good idea, after all, ignoring a caution is also dangerous - if
you don't slow down at the caution then you probably can't stop for a stop
signal. In hindsight it maybe wasn't the best choice - because commuter train
in particular operate on very busy lines, they are almost always passing
caution (yellow) signals instead of green - hence the driver gets conditioned
to hitting the pedal every time the horn sounds. This is an example, I
believe, of where 'too much' safety is a bad thing.
it coudl be redesigned at a cost, and I assume that it will have been
considered. Since I'm not involved with that industry then I have no
information on what may or may not have been discussed. Certainly the cost of
replacing all the train mounted systems, and the track mounted units, and all
whilst keeping backwards compatibility, and doing it all with minimal
interruption to services, will have been a consideration.
It's a tough question in many situations, let me give you another example.
It's possible to make a fire alarm detect a fire much sooner by making the
sensors more sensitive - quicker detection gives earlier alarm gives more
time for people to escape which you might think was a good thing. Trouble is,
if you do that then you increase the number of false alarms, more false
alarms means that people learn to ignore it. What is the value of a fire
alarm that people ignore ?
