Metrolink Wreck

[WhoozOn1st]

Metrolink adds video cameras to locomotives

"They will also be standard equipment in 57 new 'lead cars' [aka cab cars] -- passengers cars that lead commuter trains when the locomotive is pushing from behind."

 

[jis]

In normal operating circumstances I would agree with you Jishnu, but remember that this accident wasn't quite normal. First, let's talk about normal. In the case of the Metrolink train, had the engineer sailed past the approach signal (the signal right before the red one), at a rate of speed higher than allowed by that signal, then yes PTC would have kicked in and stopped the train prior to the stop signal.
However, in this case that approach signal was just prior to a station stop. Therefore the train was already slowing down to make that stop, therefore no penalty application by PTC. Now the engineer starts back up after making the station stop. By the time he reaches a speed that would be above the maximum allowed by the restricting approach signal, he's probably already covered half the one mile distance to the signal at danger.

Once he gets above the maximum allowed speed, the PTC gives him a warning that he needs to stop/slow the train down. The engineer is given at least 15 to 20 seconds, maybe even a bit more time, to actually start slowing the train. If no action is taken, a loud blaring alarm goes off that would basically wake the dead, warning the engineer that he needs to take immediate action. And again, the engineer is given several seconds, probably 15 to 20 again. By this point, the train has now probably covered another 1/4th of a mile, if not more.

PTC finally has enough of the inaction and it applies the brakes, but with less than 1/4th of a mile left between the switch and the engine, there is no way that this train is stopping before running the switch. UP coming at 40 MPH will slow once its PTC warns the engineer that there is a problem ahead, but again, he's not stopping before colliding with the Metrolink train. Yes, it won't be an 80 MPH collision, but there will almost certainly be a collision.

All that said, if PTC had been there that day, it is most likely that Mr. Sanchez would have dropped his cell phone and stopped the train in time. But, if Mr. Sanchez had been somehow incapacitated, it is unlikely that PTC would have prevented an accident. It would have mitigated the consequences of a collision, but I suspect that one still would have happened.

Alan, A properly designed PTC system enforces the invariant that the engine does not pass a signal at danger. If necessary it overrides all these niceties of blowing horns and giving 15 seconds to the engineer to react. So I still contend that any PTC system with continuous monitoring worth its salt must have been able to prevent this accident. Otherwise it ain't a real PTC system.

The job of a properly designed PTC system is to enforce the braking curve so that the train comes to a standstill before it reaches a signal at danger, with or without the cooperation of the operator. Once the braking curve is violated the operators is essentially out of the loop. Have seen such happen on a TGV. It is different from a penalty brake application. It is enforcement of the computed braking curve to stop at a particular point (or bring the speed down to some specified value by a particular point) It is impressive how smoothly it works. Now of course if one is using something on the cheap that does not deal with brake curve enforcement then all bets are off, and then it is not a system that can guarantee positive train separation.

 

[WhoozOn1st]

Here's a different, and somewhat more detailed, take on Metrolink's installation of cab cameras. From RailwayAge:

L.A. Metrolink pioneers inward-facing locomotive cameras

"All cameras and audio devices in each of the system's 52 locomotives will be activated between October 7 and 12. LDVRs [Locomotive Digital Video Recorders] will also be included in the 57 new Crash Energy Management-equipped lead passenger cars that Metrolink will place into service in the coming year."

I may have been aboard one of the new cab cars for a run from Lancaster to LAX on the Antelope Valley Line last Saturday (train 266); definitely a different design than seen/ridden previously (my first time on AV Line). An extra compartment in front (push mode) beyond main seating, and seatbacks had greater recline angle. Just not sure what "in the coming year" means, and whether some might already be in service.

Return run on train 269 included usual cab car in consist.

EDIT: Redesign of the cab cars came as a result of Metrolink's 2005 Glendale wreck, which killed 11.

 

[WhoozOn1st]

L.A. Times reader responses to the above story about Metrolink cab cameras (copy/paste to spare AU readers the slog through 11 letters on other topics):

Re “Metrolink adds video cameras to locomotives,” Oct. 6

The horrible and entirely preventable Chatsworth Metrolink collision should remind us all that rail safety is a top priority.

The threat by the Brotherhood of Locomotive Engineers and Trainmen to sue to remove cameras shows that rail crews are not on board yet with the seriousness of the rail safety problem. Neither is Gov. Arnold Schwarzenegger, who wants us all to believe in a huge construction project for high-speed trains by 2025 using expensive bonds. He needs to start thinking about improvements in trains and public safety that can happen next year.

__________________

During the last several years, my wife and I have traveled throughout Europe using a variety of public ground transportation -- trains, buses, streetcars.

We never observed a driver texting or using a cellphone. The drivers take their job seriously.

Here, the union's response was predictable. Cameras an invasion of privacy? This is public transportation. You have no rights to privacy.

 

[DET63]

Metrolink adds video cameras inside locomotive cabs

The Brotherhood of Locomotive Engineers and Trainmen, the union representing the Metrolink engineers, has called the cameras ineffective, and an invasion of privacy. 

The union plans to take legal action to remove them.

 

"If guys fall asleep or go through a red signal, whether the camera is there or not doesn't really make a difference. The camera doesn't have a magic button to stop the train," said Timothy Smith, chairman of the union's California legislative board.

 

Instead, Metrolink should staff each locomotive with two engineers and install positive train-control systems, he said. Those high-tech braking systems use digital communications with Global Positioning System technology to monitor train locations and speeds, and if engineers fail to comply with signals, the train automatically applies the brakes.

 

Metrolink board member Richard Katz said the agency plans to install the on-board automatic braking systems on all its trains by 2012.

 

[MattW]

I can't see what's wrong with the cameras. It's not a bedroom, I assume the bathrooms won't have cameras in them (if they exist) so what's the big deal? I get the feeling complacency with the rules is partially what led to this crash. "I've never crashed before while texting so it won't hurt this time." Complacency kills, and perhaps when people performing critical jobs start getting slapped with disciplinary action for rule violations, this will stop. No, there will probably never be a definite "he wasn't texting and because of that he saved the the train" incident, but prevention can sometimes be the best cure (though don't get me started on the WHO and H1N1).

 

[WhoozOn1st]

Same railroad, different wreck...

Metrolink's tab for 2005 wreck is at $30 million

"Claims arising from the Chatsworth crash are in early stages. Plaintiffs' attorneys in that case said they were planning on having their experts review the crash site early next month. Attorneys have said those lawsuits will probably overwhelm a $200-million cap that Congress imposed 11 years ago on a railroad's liability in any one accident."

 

[WhoozOn1st]

Another red-light probe for Metrolink

"Metrolink is investigating a new safety breach on the regional commuter line involving an engineer who allegedly failed to stop at a red signal last week as another train approached.

"The incident, which also is being reviewed by federal regulators, is the fourth alleged red-light violation since the Sept. 12, 2008, crash in Chatsworth that killed 25. Investigators have attributed that crash to an engineer running a red light."

Also...

"The current probe will be one of the first to incorporate footage from just-installed video cameras that are trained on engineers."

 

[Rob_C]

How about a relatively simple IR system that works like your remote control? The red light sends a signal in the direction of an oncoming train, the train gets the signals and plays a loud "RED" sound or voice that the engineer hears. I don't know what could be much better than that other than complete computer control which probably has it's own pitfalls.

 

[transit54]

How about a relatively simple IR system that works like your remote control? The red light sends a signal in the direction of an oncoming train, the train gets the signals and plays a loud "RED" sound or voice that the engineer hears. I don't know what could be much better than that other than complete computer control which probably has it's own pitfalls.

There's two problems with this approach:

1) If there isn't line of sight between the two trains. IR needs line of sight to work effectively and reliably. If there was line of sight, the crews would probably already be putting their trains into emergency.

2) Other trains - consider a double-tracked line. If a train is heading in one direction, it's going to be broadcasting IR signals to trains heading up the the other track. Sure, it could be programed to identify what track it is on, but trains routinely switch tracks. So it wouldn't be failsafe, as it would need to be updated by the crew (if you invested in the technology to make it automatic you're going to be very close to a PTC system anyway).

 

[Ryan]

He's not talking about the trains sending out the IR, but the red signals. That's actually a pretty good idea.

 

[transit54]

He's not talking about the trains sending out the IR, but the red signals. That's actually a pretty good idea.

Ah, now I understand. That is a good idea. Except, would it present a problem with double tracks and two trains moving in the same direction, one which has a red and one which has a green? There would need to be a way to limit the range of the IR signal, or target it to a particular train. Not that it couldn't be accomplished...

 

[amtrakwolverine]

what they could due is have a thing to trip the trains breaks like they due for subways. when a signal is red a lever is up and hits a valve switch on the underside of the car dumping the air.

 

[MattW]

Use a low power IR diode laser mounted track-centerline, or off the end of a curve pointed track centerline.

 

[AlanB]

The problem with anything IR based is that by the time a train going 70 MPH is within range of the IR, nothing is going to stop that train from running that red signal, short of a brick wall. And it would need to be a very thick brick wall, which would have other repercussions.

Yes, the if engineer obeyed the approach signal before the red signal, then the IR reminder might be helpful. But if the engineer isn't paying attention at all, and is still rolling along at higher speeds, then any warning triggered by an IR device will come too late.

The same problem would come into play with any trackside trip arm, by the time the train hits that arm and has its air dumped, it will be well beyond the signal and most likely in the path of another approaching train.

 

[user 1215]

Another problem would be that the second a bug splattered on the IR receiver, it would render the system useless.

 

[WhoozOn1st]

Metrolink's CEO could be moved aside

"David R. Solow has been under pressure since Chatsworth crash"

"...[Metrolink] board members said Solow appeared overloaded at times by the demands of the job, which intensified substantially after last year's head-on collision, which left 25 dead and 135 injured."

Service cuts are also on the agenda:

"To avoid or minimize the 6% fare boost initially proposed by Solow's team, the board today will consider cutting dozens of lower ridership trains used by thousands of riders each month."

 

[WhoozOn1st]

Metrolink's CEO accepts demotion

"David R. Solow, the five-county Metrolink agency's chief executive for more than a decade, agreed to step down and take a new advisory position guiding deployment of a complex safety system [positive train control] intended to prevent a repeat of last year's crash, which left 25 dead and 135 injured."

Service and fare hike questions were kicked down the road:

"Also on Friday, the board, in a sharply divided vote, put off decisions on the latest 3% proposed fare increase and possible cutbacks of dozens of trains until early next year.

"The action could exacerbate a multimillion-dollar shortfall in the current year's operating budget, officials acknowledged. But a slim majority of members said they wanted to continue exploring other cost-cutting measures and potential sources of new revenue."

FWIW, the last Metrolink fare increase reduced the Amtrak/Metrolink price difference to the point where I now prefer to pay a little extra for a Surfliner to L.A. and get the AGR points; more convenient schedule, too.

 

[WhoozOn1st]

Metrolink got a waiver to not post simple safety signs (clumsy headline!)

"Known in railroad parlance as 'delay in block' warnings, they are designed to remind locomotive engineers of speed restrictions required after they stop at passenger depots located within a series, or 'block,' of signals.

"Generally, those restrictions require trains leaving the station to stay under 40 mph and be prepared to stop until engineers see a signal indicating it is clear to go faster.

"The federal requirements were prompted by deadly, back-to-back 1996 passenger rail accidents in New Jersey and Maryland that were similar to the Chatsworth crash. In both cases, engineers failed to stop at junction signals."

___________

Does anybody recall the particulars of the N.J. and Maryland wrecks?

 

[Long Train Runnin']

From this list

The NJ crash:

February 9, 1996 – Secaucus, New Jersey, United States: New Jersey Transit Train 1254 traveling to Hoboken Terminal via the Bergen County Line runs a red signal a few hundred yards east of where the Secaucus Junction Station is currently located and slams into NJT Train 1107 traveling to Suffern, New York via the NJT Main Line. Both engineers and a passenger on Train 1254 are killed.

The MD crash:

February 16, 1996 – Silver Spring, Maryland, United States: The engineer of a MARC commuter train bound for Washington Union Station, either misses or ignores a stop signal and collides with outbound Amtrak train no. 29, the westbound Capitol Limited. The crash left 3 crew and 8 passengers dead aboard the MARC train.[65] Three die of injuries suffered in the impact; but the rest are killed by smoke and flames, the fire which may have been ignited by oil pot switch heaters. This accident lead to the FRA instituting the Delay in Block Rule, and also was a major impetus for the Passenger Equipment Safety Standards regulation (49 CFR Part 238).

 

[WhoozOn1st]

This article is posted over at the Commuter rail forum because its main focus is Metrolink's new crash-resistant cars, set to arrive from S. Korea. They were designed and built as a direct result of the fatal (11 dead) wreck at Glendale years before the Chatsworth calamity. It's posted here as well because it also touches on the subject of the Chatsworth-related story posted above yesterday.

Metrolink to get updated rail cars

"Responding to a Times report on Thursday, officials said they would review a 13-year-old federal waiver that allowed Metrolink to forgo installing simple safety signs intended to avoid crashes like last year's head-on collision in Chatsworth that left 25 dead and 135 injured. The signs were intended to remind engineers to proceed slowly, watching for signal status after they stop at stations that are between trackside control lights."

____________

The photo that accompanies the L.A. Times story shows a locomotive sporting Metrolink's new paint scheme and logo. In the wake of the cornfield meet at Chatsworth I've often thought the old (current) logo is pretty unfortunate:

To Long Train Runnin': Thanks for the info on the NJ and MD wrecks, but I asked does anybody RECALL them. Despite your advanced years I seriously doubt you have any direct memory of those 1996 accidents.

 

[Green Maned Lion]

I remember one of them. Our house keeper was on a train behind one of them and was late to work as a result.

 

[George Harris]

To Long Train Runnin': Thanks for the info on the NJ and MD wrecks, but I asked does anybody RECALL them. Despite your advanced years I seriously doubt you have any direct memory of those 1996 accidents.

Recall, as in remember when they happened? 1996 is not that long ago, so of course is the answer.

If you want the details, go to the NTSB web site and look up the accident report. These things are exhaustive in their details.

 

[George Harris]

Following are the report numbers from the NTSB web site

Title: Collision and Derailment of Maryland Rail Commuter MARC Train 286 and National Railroad Passenger Corportation AMTRAK Train 29 Near Silver Spring, MD February 16, 1996 NTSB Report Number: RAR-97-02, adopted on 6/17/1997

NTIS Report Number: PB97-916302

Title: Railroad Accident Report Near Head-on Collision and Derailment of Two New Jersey Transit Commuter Trains Near Seacaucus, New Jersey February 9, 1996

NTSB Report Number: RAR-97-01, adopted on 3/25/1997

NTIS Report Number: PB97-916301

Title: Railroad Accident Report Collision of Washington Metropolitan Area Transit Authority Train T-111 with Standing Train at Shady Grove Passenger Station, Gaithersburg, Maryland January 6, 1996

NTSB Report Number: RAR-96-04, adopted on 10/29/1996

NTIS Report Number: PB96-916304

I have thrown in the WMATA accident for flavor.

These things have both a summary posted and a pdf of the report.

 

[WhoozOn1st]

To Long Train Runnin': Thanks for the info on the NJ and MD wrecks, but I asked does anybody RECALL them. Despite your advanced years I seriously doubt you have any direct memory of those 1996 accidents.

Recall, as in remember when they happened? 1996 is not that long ago, so of course is the answer.

If you want the details, go to the NTSB web site and look up the accident report. These things are exhaustive in their details.

Thanks, I will take a look at the NTSB stuff. It was really just a joke about Long Train's age. He was about 4 in 1996.