[SystemSafety] Lac-Megantic disaster

[Peter Bernard Ladkin]

For a longer, more detailed account of some features of the Westinghouse-derived air brake system, 
people might like to look at https://en.wikipedia.org/wiki/Railway_air_brake   and 
http://www.railway-technical.com/air-brakes.shtml

Some people are likely comparing mentally with a simple vacuum-brake system. The Westinghouse-type 
air brake system has the advantages during use (to brake a running train), amongst others, of
* being more responsive (for example, differential air pressure may be chosen by the designer for 
most effective mechanical properties, whereas vacuum is limited to 1 bar differential, obviously);
* more effectively checked and maintained (it is easier to notice or find a leak in a 
positive-pressure line, for example by walking past it, than it is to find a leak into a vacuum 
line. Most train-dispatch procedures in the US according to Wiki involve a walk-past of each car to 
ensure air brakes are functioning. This would be much less effective or simple with vacuum-operated 
brakes);
* lighter (lower-pressure-differential systems are generally heavier than higher-pressure 
counterparts in all sorts of engineering contexts).

Stopping a running train effectively is - I would say obviously - a different problem from 
maintaining a stationary (parked) train stationary. The first changes the dynamics of a situation 
and there are effects such as take-up of slack (conversely, slack increase on release-of-braking 
command) which make the dynamic situation quite complicated, I should think, and the system issues 
are thus much wider than a single wagon. Whereas the latter only involves keeping each individual 
wagon where it is, and thus invites a unit solution. I would guess that having separate systems for 
each task eliminates the need for compromise, keeps the reliability issues separate, and likely 
reduces technical complexity.

The question which arises with the Lac-Megantic disaster is why both systems should be considered 
needed to maintain the parked train stationary. The answer is presumably: the train is parked on a 
gradient. Another answer may be that an engineer finds it easier to leave an engine running to power 
the air brake system than to go down the train setting individual hand brakes. Indeed, this second 
answer has been hinted at publicly by the head of the railroad company involved. We'll see what 
transpires with that.

I'd be keen on hearing thoughts on the practicality of a requirement to park trains only on level 
ground in NA. That would eliminate the risk of "migration to the boundary" procedural issues with 
engineers trying to save themselves a bit of work as suggested above.

Obviously, if you have a longish train, bits of it might be on level track and bits not, and it 
might be hard to tell (one might have to survey and signpost parking spots with allowed train 
lengths and position). I read somewhere that the gradient between Nantes and Lac-Megantic is 1.2%. 
Can anyone confirm?

PBL

On 7/12/13 12:51 AM, Matthew Squair wrote:
> Like all systems there's an operational context that goes hand in hand with safe operations. In this
> case the system is designed so that if you lose pressure, due to a burst hose say, the system will
> 'fail safe' and stop the train. Once stopped you can apply the park brakes to hold you on the grade.
>
> The pneumatic train brakes are not (emphasise not) designed to be used as a means of holding the
> train while it's parked. Just like your car, there's a mechanical park brake. And just like your car
> you shouldn't hop out and walk off without applying the park brakes.

Prof. Peter Bernard Ladkin, Faculty of Technology, University of Bielefeld, 33594 Bielefeld, Germany
Tel+msg +49 (0)521 880 7319  www.rvs.uni-bielefeld.de