Head on collision Chicago 30th Sep 2013

[WatcherZero]

33 injured in a head on collision between an out of service train and an in service train at Harlem Station.

33 people have been injured in Chicago as an out-of-service train collided head-on with another train stopped at Harlem Station at around 8am this morning local time (13:00 GMT).

The Chicago Transport Authority are investigating how the collision happened and why the out-of-service was on the track, but said that there were luckily no serious injuries, with 33 people taken to nine different hospitals.

The Blue Line service was initially suspended between Forest Park and Kedzie Avenue with shuttle buses provided, but has since resumed service. However, trains are not currently stopping at Harlem station.

CBS are reporting that police are investigating that the out-of-service may have been hijacked, but we do not yet have any evidence to support this claim.

http://descrier.co.uk/world/2013/09...ampaign=us-train-collision-chicago-injures-33

The suspicion that the train was hijacked comes from the fact that it was shortformed of only four carriages while most trains in the peak are longer, no other evidence to suggest hijacking so far. Its reported the train was under maintenence in the rail yard and to reach the station it would have had to pass two junctions and a driver alarm, collision speed estimated to be 25mph.

 

[Murph]

The suspicion that the train was hijacked comes from the fact that it was shortformed of only four carriages while most trains in the peak are longer, no other evidence to suggest hijacking so far. Its reported the train was under maintenence in the rail yard and to reach the station it would have had to pass two junctions and a driver alarm, collision speed estimated to be 25mph.

Pure speculation here, but offered to mitigate any possible t*rr*r*st hysteria, or similar (since that's what Joe Public may take from "hijacked")…

Parking brake not set / defective / being worked on. Air leaks from brake cylinders over time (perfectly normal, and expected) causing service brakes to release (as expected). Either trailing points or bad luck in how the points were set, and the right combination of gradients. No air will mean deadman system inop. IF that all fits the circumstances, no human involvement required (other than possible negligence involving the parking brake).

Or, yeah, it could have been some numpty that wanted to "play trains".

Or, could be they have sub-contracted Arriva's Swedish cleaners in…

 

[tsr]

The Ghost of DCR... the "Stafford Spectre"?

In seriousness, I'm glad there were no serious injuries/fatalities to add to the toll across the pond. Whilst all networks have their many differences, our friends in North America in general have had a few nasty accidents of late.

 

[b0b]

ex-pat here that lives really close this accident...

there are some other theories on why it might have been driven by an unauthorized person:

- the train was not in service, it was awaiting transfer to a depot, so should have had brakes set and left alone

- it apparently managed to cross at least 2 interlocks. CTA trains use a similar system to the LU tripcock system to prevent SPADs. I wonder if the CTA system doesnt' work if the train is traveling the wrong way though.

- the train has a deadmans handle that needs to be kept depressed, I also think these trains use some sort of vigilance device as well

- the Blue Line features ATP that should have kicked in to prevent the train passing reds. Anyone that has been on the Blue Line is familar with the ATP as it gives speed warnings ALL the time. I believe you can override the ATP pass reds & the tripcock system at slow speeds though (want to say under 15mph)

its all very strange that all of these saftey systems somehow were defeated....
--- old post above --- --- new post below ---

Either trailing points or bad luck in how the points were set, and the right combination of gradients. No air will mean deadman system inop.

Theres a fair uphill gradient prior to the Harlem station. It's not a big one, but not insignificant if the train had "run-away".

 

[Murph]

its all very strange that all of these saftey systems somehow were defeated....

Without intimate engineering knowledge of these particular trains, making some assumptions based on common railway technology, the one thing which seems to fit everything described perfectly, for me, is no air combined with no/loss of parking brake.

Tripcocks, ATP, deadman's handle, vigilance, service brake, emergency brake, etc, they all have one thing which they typically rely on. The air brakes. Take away the air pressure (and it's entirely normal for it to be completely lost when a train has been parked for a while), and none of those systems will work. This is not something new, it's been like that "forever", long before most modern safety systems were conceived. Additionally, depending on train design, many (all?) of those driver safety systems will only be active when a driving cab is properly enabled (by design, tripcock, deadman's, vigilence, and similar must all be inhibited in inactive cabs).

There's also not really much scope for practical improvement to the system, as air brakes are proven to be extremely safe, fail-safe, reliable, and practical under all normal circumstances. The one thing you absolutely must do, however, is ensure that a parked/shutdown train is properly secured by something other than the air brake (e.g. mechanical parking brakes, chocks, chains, etc, whatever local rules and circumstances call for). The correct assumption for a parked/shutdown train is that the air brake system will be 100% unreliable/ineffective, and that it will only become reliable and effective again after the system has been fully charged with air.
--- old post above --- --- new post below ---

Theres a fair uphill gradient prior to the Harlem station. It's not a big one, but not insignificant if the train had "run-away".

That's something which only detailed local knowledge can properly address. If the highest point on that gradient is below the point where the train was parked, it becomes quite possible for it to coast over it. If it is higher, I'll readily and gladly label the incident as extremely strange, unless there's something else to explain additional energy being added to it (e.g. shunting mishap).

 

[b0b]

If you have any interest, here are some accident reports from previous Blue Line incidents, that talks about the ATP system

http://www.ntsb.gov/doclib/safetystudies/SIR0201.pdf

I'm beginning to wonder if your theory is good, its kind of a difficult environment to judge all the gradients west of where the Blue Line dives under the CSX freight line then cross es an interstate which is in a cutting but then rises I think to the Forest Park station and on to the depot. The depot certainly is flat though so I wonder how the train got moving in the first place, if it did run away.

The train was reported out of use for a few days and had faults in 3 of the 4 cars.

 

[Murph]

If you have any interest, here are some accident reports from previous Blue Line incidents, that talks about the ATP system

http://www.ntsb.gov/doclib/safetystudies/SIR0201.pdf

I had a fairly quick speed-read of that, thanks for the link. Nothing immediately jumped out at me to invalidate my hypothesis.

I'm beginning to wonder if your theory is good, its kind of a difficult environment to judge all the gradients west of where the Blue Line dives under the CSX freight line then cross es an interstate which is in a cutting but then rises I think to the Forest Park station and on to the depot. The depot certainly is flat though so I wonder how the train got moving in the first place, if it did run away.

That, and "what happened to the parking brakes?", are the $64k questions! The other thing is whether it really is perfectly flat? There are several historical incidents of runaway locos eastbound out of Edinburgh Waverley from tracks which visually appear to be flat, but there was sufficient gradient to allow them to start rolling. It's very difficult to properly judge by eye alone, and doesn't take much to start rolling away once the brakes are fully released. In an ideal world, the depot would have trap points on all tracks leading out onto the main line.

The train was reported out of use for a few days and had faults in 3 of the 4 cars.

Out of use for a few days certainly fits with "no air". I think that we'll probably only know if the faults are relevant once an insider speaks, or the NTSB publish. There's certainly many faults which could take a train out of service, but not contribute to this incident.

 

[185]

If it's not already been mentioned, Blue Line is an 'L' train for those in the know.

 

[Murph]

Well, it seems that my theory that it was a simple case of the brakes bleeding off, combined with no parking brake was wrong. Thankfully, I was correct about it not being "hijacked", and there doesn't seem to be any malicious behaviour involved. The NTSB have released a couple of preliminary documents which are pointing the finger at it being a quite strange incident of electrical defect/failure.

http://www.ntsb.gov/Shutdown/R-14-001-002.pdf
http://www.ntsb.gov/Shutdown/R-14-003.pdf
(PDFs attached to this post, since it seems likely that those URLs will change when the NTSB comes out of shutdown, when the legislators successfully uncouple their heads and posteriors…)

In short, it seems that there was a good 2-car unit coupled to a bad 2-car unit, and the bad unit's control systems failed in such a way that they applied traction, overpowering the brakes on the good unit. A speculative causal factor seems to be shorted wiring, and/or water ingress into the electrics:

Upon inspection one of the cars of the unoccupied train was found to have thermally damaged wiring and water in electrical connection boxes on the car. CTA has begun inspecting its fleet of similar cars to determine if similar conditions exist on other cars.

 

[b0b]

Still it is bizarre that a train is able to "reset itself" from an emergency brake application.... and keep doing it! thats a safety critical flaw.

 

[Murph]

Still it is bizarre that a train is able to "reset itself" from an emergency brake application.... and keep doing it! thats a safety critical flaw.

The real safety critical flaw is the damaged wiring or water ingress to wiring which caused the traction to engage, not how the normal safety systems operated. It seems likely (but this is speculative) that there was no master key in any cab desk, and the reversers were at neutral or off. In that situation, it's likely that some or all of the normal driver safety systems are inactive.

Traction engaged with no active cabs is unlikely to be something which the safety systems are designed to deal with, as it's a hugely abnormal situation, well outside what one would expect for the design parameters. The control systems are supposed to not allow that to happen in the first place.

If it was a car, the equivalent would be approximately the gearbox engaging drive, despite the shifter being in park, then giving itself sufficient throttle to overpower the brakes!

 

[Yew]

If it was a car, the equivalent would be approximately the gearbox engaging drive, despite the shifter being in park, then giving itself sufficient throttle to overpower the brakes!

I guess you drive auto?

 

[Murph]

I guess you drive auto?

I suppose I should have specified auto, although that is what I prefer to drive, given the choice. It also seemed like the appropriate comparison to a train.