In article , Nick Evans
wrote:

"Michael Bell" wrote in message
...
Derailment is the classic railway and tramway disaster.
Remember the derailment a few years ago in North Germany? The flange
of a wheel broke, the wheel came off the track and bounced along the
sleepers. The coach lurched horribly, the passengers looked at each
other, but they didn't feel able to pull the handle. The wheel jumped
back onto the track and ran along for another kilometre or so, then
came off the rail again and this time the coach slewed sideways and
hit the pier of an overbridge and the following coaches rammed into
it. About 80 killed. Could a derailment detector have prevented this
accident or reduced its severity? Does such a device exist? Presumably
the resultant action would be to put the brakes on automatically. Or
only slam the brakes on on the vehicles BEHIND the derailment?
That's more difficult to get right, especially as railway vehicles
tend to be 2-directional.

There seem to me to be the following possibile ways you might
detect derailment :-

* Too-fast angle increase. Fit a gonoimeter (yes, that's the
correct word for an angle-measurer!) between car and bogies and too
fast a rate of angle increase signals derailment.

Can we be sure it would always signal an accident?

* Vibration/Jolts. Fit accelerometers on axle boxes, too great
a jolt signals a derailment.

Can we be sure it would always signal an accident? But it
might also indicate serious trouble like a broken wheel BEFORE a
derailment, and that would be useful.

* Electrical continuity between pairs of axles. Using a
toroidal transformer wound round the axle (or coils facing the wheel)
put a voltage on one axle and measure the return current in the other
axle.

* Detect the presence of the rail using magnetic switches
("proximity detectors") placed between the pairs of wheels of a bogie.

This is an established technology, of good reliability and
cheap. So far it is my favourite.


Any safety device costs :-

* Money

* Weight

* Space

* Maintenance effort. Poorly thought-out safety devices can
cause more danger than they prevent, and a device like this should
have a lower failure rate than the danger it is designed to prevent.
Derailment is a very rare event on plain track, though it is more
likely at points and tram track is liable to rubbish or vandalism. But
even so derailment is rare. Can any device justify the expense? Should
the effort instead go into ensuring that the track and running gear
are in good order?

Where might I find further information on derailment?

--

Michael Bell

I would guess that some sort of visual detection system could work. I was
recently playing with this new addon 'toy' for the playstation2 which uses a
camera/image reconition to play games. But even for what is essentially a
toy... the image reconition is very good. Now picture (no pun intended) a
camera pointing to each axel or wheel and when something that does not match
a pattern or soemthing happens. It can either slam the brakes on or at least
alert the driver who can look at the camera image from a screen in the cab
and decided weather to stop or not.


Regards
Nick Evans



That is a possibility I hadn't thought of. It has it's problems, for
example mud being thrown up onto the lens, lighting at night, expense,
complexity and bulk. I still prefer the proximity detector, which is a
minaturised version of the loop detectors you get before traffic
lights. But thank you for your thought.


Regards

Michael Bell

--