On the Train Wreck in Quebec

News reports on this are still quite confusing,  but here is what I have been able to determine as pretty much the facts:

The train was parked and "secured" near Nantes (uphill and 7 miles away from Lac-Megantic),  and the driver went to a hotel for the night.  In this instance,  "secured" seems to mean a running diesel engine powering the air brake system,  I'd guess by the two-pipe system,  such that the air brakes were set in the locomotive and all train cars. 

Then there was some sort of fire in the parked locomotive,  to which the Nantes fire department and a railroad engineering division person responded.  They put out that fire,  which seems to have been fairly minor,  and shut down the running engine.  They had a procedure to follow,  and they followed it.  Everybody went home.

About an hour later,  the entire train rolled away,  picking up speed gravitationally,  and entering Lac Megantic fast enough to derail and cause the disaster. 

This is my suspicion,  based on what I have read about train air brakes: 

When they shut down the locomotive,  they killed the air compressor.  Train car second pipe pressures then fell due to air leakage,  which is inevitable.  Reduced train line pressures would act to apply car brakes,  but not with leak-drained reservoirs!  The second line that keeps the reservoirs filled would have been off,  too,  with the locomotive shut down. 

The locomotive air brake would also have been disabled by bleed-down with the compressor off.  I'm not sure about sequence,  but sooner or later,  all of the air brakes would have failed from bleed-down.  Apparently this took about an hour. 

Apparently,  nobody thought to set at least one car's handbrake,  and this must not have been in the locomotive fire procedure that they were following in Nantes.  One or two cars' handbrakes could have secured the train,  even without any compressed air at all in the air brake system.  That could have prevented this disaster.

As for the tank cars themselves,  these were unpressurized-liquid cars,  and reported to be DOT-111 designs,  which are well-known to be thin-skinned and easily punctured.  They were hauling a light crude that comes from fracking operations in North Dakota.  It would have flammability characteristics closer to diesel than heavy fuel oil. 

The problem with closed flammable liquid tanks is overpressurization explosions when exposed to fire.  Those are not detonations like high explosives,  but they are still extremely violent.  Once the cars are thrown together in a wreck and some torn open,  the fire starts.  Explosions become inevitable. 

The disaster is better and easier to prevent,  than to fight after-the-fact.

Recommendation:  set some handbrakes if you shut off the locomotive on a parked train.  Leave a note to that effect in the cab before you go. 

GW

Second thoughts (later,  same day):

Being a freight,  this was most likely a one-pipe air brake system,  not a two-pipe.  It doesn’t matter to the outcome.  Turn off the air compressor (and they did),  and after a while,  every brake component bleeds to zero.  Once that happens,  brakes release. 

When the last brake releases (car or engine,  doesn’t matter),  the train is free to roll,  unless some mechanical handbrakes somewhere were set.  Once free to roll,  if parked on any sensible slope at all,  the train will roll away under gravity.  That is inevitable. 

One pipe system or two,  same outcome as already described.

Same preventative as already described:  set a handbrake or two,  more of them on a steep slope.  There’s plenty of time to release them while the cars’ air brake components “charge up”,  with the (running) locomotive brake set to hold the train. 

 Update 7-19-13:

It is now my understanding that the engineer was supposed to have set 10 or 11 of the handbrakes on this train.  Obviously,  this didn't work. 

There are then only 3 possibilities:  (1) he did not set them,  (2) he did set them,  but they were ineffective (for a reason of considerable interest),  or (3) somebody else released the handbrakes. 

I think the authorities have their work cut out for them,  finding out which of these three possibilities caused the disaster.  It is easy enough to blame the engineer,  but there are two other possibilities that must be eliminated before that is a credible action for anyone to take.

Neither of the other two possibilities is a very comfortable thought.  Yet,  they MUST be dealt with.  If they are not,  then any "final report" on this disaster is neither credible nor useful.