Trains.com

DC Metro Collision

15323 views
125 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    August 2005
  • From: At the Crossroads of the West
  • 11,013 posts
Posted by Deggesty on Thursday, June 25, 2009 11:03 AM

Paul_D_North_Jr
Even if only a quarter-mile's distance and visibility was available to stop, if the operator reacted promptly by then the train should have stopped, or been moving much slower than it apparently was.

We cannot know how long it took the operator to react; could it not have taken longer than two seconds?

Johnny

Johnny

  • Member since
    November 2006
  • From: Southington, CT
  • 1,326 posts
Posted by DMUinCT on Thursday, June 25, 2009 10:26 AM

Phoebe Vet

selector

I'm going to stick my neck in here and state that what you pay a person and how they orient themselves to their work are very poorly related.  Recent history of the work and moral ethics of people at all social and financial strata will support me.  Was the texting engineer from last fall paid minimum wage?  Was Madoff paid minimum wage?  My wife works at a slightly above minimum wage job at a popular coffee chain up here in Canada, and I can asure you that she brings her all to her work because that is her nature.

We should stick to sensible premises.

-Crandell

Crandell:

I don't think the contempt that you read in that "minimum wage" comment was directed at the motorman so much as at the employer who feels that the operator monitoring the computer is worth less than the one who is actually manipulating the controls.

The wage paid for a given job does not have much of an effect on the employee's motivation.  It does, however, seriously impact the quality of people who apply for the job when it is vacant.  When the pay is low, you get few applicants and often must hire some people that you wouldn't otherwise.  You will get some good ones, but they will probably always be looking for something better.  When you pay well, you get many applicants and you can be very fussy about which one you hire.

Sometimes low wages are all that can be offered because there is not enough markup in the end product to support higher.

I have no idea of the pay scale of a Subway Motorperson.  In Washington DC it must be a Union Scale.

As for up here in Boston, Commuter Rail is B.L.E./United Transportation/Teamster Union.  Locomotive engineers who work a split shift (morning rush hour - afternoon rush hour) can make more than $100,000 a year.  With there F40PH or GP40MC, they drive 6 and 8 car trains at 80mph.

Don U. TCA 73-5735

  • Member since
    October 2006
  • From: Allentown, PA
  • 9,810 posts
Posted by Paul_D_North_Jr on Thursday, June 25, 2009 10:26 AM

At a deceleration or braking rate of -1.0 MPH per second [-1.5 ft./ sec. per sec.] - which is a very slow figure, almost unbelievably so - I figure the time to stop at 45 seconds and the distance at about 1,520 ft. after the brakes are applied [D = 1/2 x Decel. Rate in ft./sec. per sec. x Time squared, in secs.], plus the distance traveled during the operator's perception and reaction time.  If we allow 2 seconds for that at 45 MPH = 66 ft. / sec. for 132 ft., then the total time is 47 seconds and the total distance is about 1,650 ft. = 0.31 mile.

At a decel rate of 2.0 MPH per second - still pretty slow - the braking time is 22.5 secs., and the distance is about 760 ft., for a total time of 24.5 secs. and a total distance of about 890 ft. = 0.17 mile.

At a decel rate of 3.0 MPH per second - more realistic - the braking time is 15.0 secs. and the distance is about 505 ft., for a total time of 17 secs. and a total distance of about 640 ft. = 0.12 mile.

Even if only a quarter-mile's distance and visibility was available to stop, if the operator reacted promptly by then the train should have stopped, or been moving much slower than it apparently was.

- Paul North.

"This Fascinating Railroad Business" (title of 1943 book by Robert Selph Henry of the AAR)
  • Member since
    September 2007
  • From: Charlotte, NC
  • 6,099 posts
Posted by Phoebe Vet on Thursday, June 25, 2009 9:46 AM

At 45 mph it would take 20 seconds to travel a quarter of a mile with no braking.

Dave

Lackawanna Route of the Phoebe Snow

  • Member since
    April 2001
  • From: US
  • 1,103 posts
Posted by ValleyX on Thursday, June 25, 2009 9:30 AM

 

20 seconds to stop in a quarter mile? Which figures out to running 45 MPH and coming around a curve and reaction time, believing you're seeing what you shouldn't be seeing and taking action accordingly.  I've no idea but some of you might know what the stopping distances are on light rail trains such as this, a quarter mile isn't much 

Oh, and as for Casey Jones, yes, he did have quite a record but he was a product of his times, too. 

  • Member since
    March 2005
  • 40 posts
Posted by harpwolf on Wednesday, June 24, 2009 11:10 PM

We in San Francisco are watching this with keen interest, as our BART system also uses ATO (automatic train operation. 

 

Media is saying the lead car in the second train was a "B" car.  That doesn't sound right to me at all, I thought B-cars did not have a cab or controls adequate for more than shunting. Anyone know more?

  • Member since
    April 2008
  • 123 posts
Posted by Jerry Pier on Wednesday, June 24, 2009 6:22 PM

My saying LTK designed the cars is misleading. They wrote the specm as did PBQ&D, for Bart but Rohr was responsinble for designing a car that met thoise specs in each case.

With respect to the Marta cars, I was involved in that contract only briefly but it could well have been. have been LTK.

Jerry Pier

JERRY PIER
  • Member since
    October 2006
  • 1,123 posts
Posted by HarveyK400 on Wednesday, June 24, 2009 5:04 PM

As an afterthought, I recall reading a long time ago about concerns for cross-talk between tracks with audio-frequency carrier signals.  This may be a possible explanation for a false clear signal in the absense of the shunted circuit of the occupied track.

  • Member since
    September 2008
  • 1,112 posts
Posted by aegrotatio on Wednesday, June 24, 2009 4:03 PM

 The operator had very little to do with the accident, though there is some question because it appears she had at least 20 seconds to stop the train after seeing the stopped train about 1/4 mile ahead of her.

The evidence is leaning towards failure of wayside equipment to provide control data to the train, or the train failing to acknowledge the control data signals from the wayside equipment.

 The arguments from NTSB about crashworthiness of the Rohr cars is total bupkis.  The object of the game is to keep trains apart, PERIOD.  The speed of this collision would have killed on any train--Rohr, Breda, CAF, Alstom.

Unfortunately we have no telemetry which is very disturbing, not just because the Rohr cars don't have 'black boxes' but also because Central Control should have telemetry transmitted to it by the wayside equipment.

Please, visit the Wikipedia page about the WMATA Metrorail system.  I ride it regularly.

http://en.wikipedia.org/wiki/Washington_Metrorail

 

 

 

  • Member since
    January 2001
  • From: Atlanta
  • 11,971 posts
Posted by oltmannd on Wednesday, June 24, 2009 2:28 PM

HarveyK400
Furthermore, how could a signal command not be shunted by the lead train in the track circuit that nonetheless had been commanded to slow and stop?

A failure of the wayside track circuit/signal system and/or on board cab signal system creating a "false proceed" is one way.  There were 6 in 2008 due to equipment failures in the US.

Take a look!  http://safetydata.fra.dot.gov/OfficeofSafety/publicsite/affp/Fpbrowse.aspx

-Don (Random stuff, mostly about trains - what else? http://blerfblog.blogspot.com/

  • Member since
    September 2007
  • From: Charlotte, NC
  • 6,099 posts
Posted by Phoebe Vet on Wednesday, June 24, 2009 1:39 PM

selector

I'm going to stick my neck in here and state that what you pay a person and how they orient themselves to their work are very poorly related.  Recent history of the work and moral ethics of people at all social and financial strata will support me.  Was the texting engineer from last fall paid minimum wage?  Was Madoff paid minimum wage?  My wife works at a slightly above minimum wage job at a popular coffee chain up here in Canada, and I can asure you that she brings her all to her work because that is her nature.

We should stick to sensible premises.

-Crandell

Crandell:

I don't think the contempt that you read in that "minimum wage" comment was directed at the motorman so much as at the employer who feels that the operator monitoring the computer is worth less than the one who is actually manipulating the controls.

The wage paid for a given job does not have much of an effect on the employee's motivation.  It does, however, seriously impact the quality of people who apply for the job when it is vacant.  When the pay is low, you get few applicants and often must hire some people that you wouldn't otherwise.  You will get some good ones, but they will probably always be looking for something better.  When you pay well, you get many applicants and you can be very fussy about which one you hire.

Sometimes low wages are all that can be offered because there is not enough markup in the end product to support higher.

Dave

Lackawanna Route of the Phoebe Snow

  • Member since
    October 2006
  • 1,123 posts
Posted by HarveyK400 on Wednesday, June 24, 2009 12:15 PM

First, older aluminum cars do not seem to be a significant issue here.  The damage seems consistent with a moderate rate of speed and the masses of both 6-car trains.  If anything, the fact that one car overrode another without more substantial damage suggests a reassuring degree of strength and structural integrity.  If the cars had been built much stronger and not overridden one another, I suspect many more passengers in all the cars would have been seriously injured or killed from being "thrown" from their feet or seats.

Secondly, I suspected, as many have written, that the malfunction occurred somewhere in the computer and ato/cab signal systems.  As far as detection, even a stopped train draws considerable power for lights and air-conditioning to shunt signal circuits.  Furthermore, how could a signal command not be shunted by the lead train in the track circuit that nonetheless had been commanded to slow and stop?

  • Member since
    December 2001
  • 8,156 posts
Posted by henry6 on Wednesday, June 24, 2009 11:52 AM

I have often seen people get big checks for a job they are not really qualified to do and people who are underpaid forthe job they are doing well.  What happens after the individual differences, etc., is a lack of interest or ability in doing better; then crowd mentality sets in and the morale of the group sinks and nobody gives a *** and the work environment worsens more.  I am in a business where it is now deemed that anyone can be snagged off the street with just a high school diploma (or not) and be have a career doing what us that have gone on before have had to have college degrees and years of experience for the industry to achieve what it has.  Now management is reinventing the wheel but they are leaving out the spokes.  Do you think morale is good, Do you think our customers are getting what they think they are getting, what they think they are paying for?  No, if you want quality and safety and integrety, you have to hire qualified people, careing people, and pay them a decent wage to do it.  Look at what sports figures get.  Isn't an employee that can earn  a company millions of dollars a year because of his experience, education, and ability, worth more than minimum wage?  A baseball, football, basketball, hockey, soccer, tennis, and golf "employees" all get more than that and what do they actually contribute to the society except maybe one spectacular "sports moment"  How about the guy who makes sure your car is solid and safe and nobody's gonna get killed because it the car is perfect? 

RIDEWITHMEHENRY is the name for our almost monthly day of riding trains and transit in either the NYCity or Philadelphia areas including all commuter lines, Amtrak, subways, light rail and trolleys, bus and ferries when warranted. No fees, just let us know you want to join the ride and pay your fares. Ask to be on our email list or find us on FB as RIDEWITHMEHENRY (all caps) to get descriptions of each outing.

  • Member since
    November 2003
  • From: Rhode Island
  • 2,289 posts
Posted by carnej1 on Wednesday, June 24, 2009 11:12 AM

BigJim

edblysard
And no, her putting the train into emergency did not prevent the accident, but I bet it saved a lot more lives than we will know.


Maybe she will be the hero of this whole thing. How odd it would be if her name just happened to be Casey Jones.

Given that Mr. Jones was known for disregarding safety rules one would hope not..

"I Often Dream of Trains"-From the Album of the Same Name by Robyn Hitchcock

  • Member since
    February 2005
  • From: Vancouver Island, BC
  • 23,330 posts
Posted by selector on Wednesday, June 24, 2009 10:57 AM

I'm going to stick my neck in here and state that what you pay a person and how they orient themselves to their work are very poorly related.  Recent history of the work and moral ethics of people at all social and financial strata will support me.  Was the texting engineer from last fall paid minimum wage?  Was Madoff paid minimum wage?  My wife works at a slightly above minimum wage job at a popular coffee chain up here in Canada, and I can asure you that she brings her all to her work because that is her nature.

We should stick to sensible premises.

-Crandell

  • Member since
    January 2001
  • From: Atlanta
  • 11,971 posts
Posted by oltmannd on Wednesday, June 24, 2009 10:45 AM

ns3010

henry6
But the question which comes to my mind is that if you have to put a person onboard to oversee the computer system...like PATCO and Metro here...why not give them control and let the computer system override them instead?

Actually, that's a good point. Unfortunately, it's probably to save money:
"We'll pay you minimum wage to open doors and prevent the train from crashing, instead of paying you five times as much to actually control the train."

PATCO does exactly what Henry suggests.  The operator runs the train manually during off peak periods with the ATC system as a safety backup.  This keeps the operators proficient.  Only when the need to squeeze out all the variablility of manual operation in order to keep and maintain rush hours schedules and headways do they run in full automatic mode. 

(there are no on board computers involved in PATCO's ATC system!  It's just cab signals, relays and magnet valves.)

-Don (Random stuff, mostly about trains - what else? http://blerfblog.blogspot.com/

Moderator
  • Member since
    November 2008
  • From: London ON
  • 10,392 posts
Posted by blownout cylinder on Wednesday, June 24, 2009 9:12 AM

ns3010

henry6
But the question which comes to my mind is that if you have to put a person onboard to oversee the computer system...like PATCO and Metro here...why not give them control and let the computer system override them instead?

Actually, that's a good point. Unfortunately, it's probably to save money:
"We'll pay you minimum wage to open doors and prevent the train from crashing, instead of paying you five times as much to actually control the train."

I've seen a few comments regarding the accident as something that could have been prevented by including the Human Being in the control of the train. But "Money" is the ruling element here----as above--

Any argument carried far enough will end up in Semantics--Hartz's law of rhetoric Emerald. Leemer and Southern The route of the Sceptre Express Barry

I just started my blog site...more stuff to come...

http://modeltrainswithmusic.blogspot.ca/

  • Member since
    November 2008
  • From: North Jersey
  • 1,781 posts
Posted by ns3010 on Wednesday, June 24, 2009 9:03 AM

henry6
But the question which comes to my mind is that if you have to put a person onboard to oversee the computer system...like PATCO and Metro here...why not give them control and let the computer system override them instead?

Actually, that's a good point. Unfortunately, it's probably to save money:
"We'll pay you minimum wage to open doors and prevent the train from crashing, instead of paying you five times as much to actually control the train."

My Model Railroad: Tri State Rail
My Photos on Flickr: Flickr
My Videos on Youtube: Youtube
My Photos on RRPA: RR Picture Archives

  • Member since
    January 2001
  • From: Atlanta
  • 11,971 posts
Posted by oltmannd on Wednesday, June 24, 2009 8:46 AM

Jerry Pier

Yes, the 1000 series were built in Winder, GA but the the design was by LT Klauder Consultants and the engineering took place in Chula Vista, CA. The body was constructed with long aluminum extrusions manufactured in Switzerland. The extrusions were "Pop Riveted" together to attain the required height dimensions. The result was a very strong and light weight car. Unfortunately, Rohr aerospace engineers designed for stress rather than deflection witrh the result that tests of the first car revealed that properly fitted doors would hang up with a crush car load. This was corrected with a steel door frame. The car passed all of the specified strength requirements.

Thanks for setting me straight!  I didn't know LTK was that heavy into transit car structural design.

The cars appear to be very similar to the first batch of Atlanta's MARTA cars.  I wonder if they are related.

-Don (Random stuff, mostly about trains - what else? http://blerfblog.blogspot.com/

  • Member since
    April 2003
  • 305,205 posts
Posted by Anonymous on Wednesday, June 24, 2009 8:46 AM

I suppose the automated train wasn't texting at the time...

If there was something on the rails (leaves?), that could have caused a failure of the stopped train to trip the block detectors, giving the second train a green signal into a block that wasn't actually empty.

  • Member since
    January 2001
  • From: Atlanta
  • 11,971 posts
Posted by oltmannd on Wednesday, June 24, 2009 8:40 AM

For any train control system, there are ALWAYS failure modes. 

Some of them are there on purpose.  For example, there has to be a way to cut the system out in order to move the train in non-controlled territory, such as into and out of a storage yard or shop.  Or, to move the train when the system fails.  There are usually rules wrapped around who, when, where and how the cut-out can be used.

There can be "false clear" signal indications as well.  Even in the "vital relay" world, there can be malfuncitons.  I am aware of one where the grease in the signal relay hardened and prevented the armature from dropping.  I am aware of another one where a bullet from a hunter's rifle prevented a searchlight signal from dropping back to stop.  Similarly, the relay valves in the braking system can stick, be cut out, have a kinked line running to them, etc.

If the Metro uses coded track circuits and on-board cab signalling - well known hardware with a long history - all it would have taken was for the "clear" decoder to stick through some highly improbable combination of failures. 

The "computer control" is likely just a "non vital" CTC overlay to keep the trains running.  The signal safety system is likely unrelated.

-Don (Random stuff, mostly about trains - what else? http://blerfblog.blogspot.com/

  • Member since
    April 2008
  • 123 posts
Posted by Jerry Pier on Wednesday, June 24, 2009 8:38 AM

Yes, the 1000 series were built in Winder, GA but the the design was by LT Klauder Consultants and the engineering took place in Chula Vista, CA. The body was constructed with long aluminum extrusions manufactured in Switzerland. The extrusions were "Pop Riveted" together to attain the required height dimensions. The result was a very strong and light weight car. Unfortunately, Rohr aerospace engineers designed for stress rather than deflection witrh the result that tests of the first car revealed that properly fitted doors would hang up with a crush car load. This was corrected with a steel door frame. The car passed all of the specified strength requirements.

Rohr also built the first two orders of the San Fransico Bart system at Chula Vista, CA, using the same construction technique. The spec in this case was by PBQ&D. To the best of my knowledge this series is still running.

At the time all this was going on, I was Engineering & Program Mgr for the Turboliner trains. It was planned to build these in Winder also but when the WMATA cars were delayed, the Turboliners stayed in Chula Vista, for which i was duly greatful. I visited Winder to assess operations there and also provided assistance to the manufacturer of the disk brakes.

Subsequent orders were built by Breda of Pistoia, Italy which had got its feet wet with a contract for ClevelandTransit System light rail cars. I was back at WABCO by this time, managing Passenger and Transit Sales and had the priviledge of negotiating the brake order with Breda.

I concur that NTSB's statement is a CYA effort. Train control and safety systems are supposed to prevent collisions such as this.

Jerry Pier

JERRY PIER
  • Member since
    December 2001
  • 8,156 posts
Posted by henry6 on Wednesday, June 24, 2009 7:53 AM

I agree there are a lot more places the computer system could go wrong from the main to the detector system to the on board equipment.  But before we totally fry this technology, look at BART and PATCO systems which have had so few glitches over so many more years...plus several European and Asian systems, too.

But the question which comes to my mind is that if you have to put a person onboard to oversee the computer system...like PATCO and Metro here...why not give them control and let the computer system override them instead?

RIDEWITHMEHENRY is the name for our almost monthly day of riding trains and transit in either the NYCity or Philadelphia areas including all commuter lines, Amtrak, subways, light rail and trolleys, bus and ferries when warranted. No fees, just let us know you want to join the ride and pay your fares. Ask to be on our email list or find us on FB as RIDEWITHMEHENRY (all caps) to get descriptions of each outing.

  • Member since
    April 2001
  • From: Roanoke, VA
  • 2,020 posts
Posted by BigJim on Wednesday, June 24, 2009 6:53 AM

edblysard
And no, her putting the train into emergency did not prevent the accident, but I bet it saved a lot more lives than we will know.


Maybe she will be the hero of this whole thing. How odd it would be if her name just happened to be Casey Jones.

.

  • Member since
    March 2002
  • 9,265 posts
Posted by edblysard on Tuesday, June 23, 2009 11:21 PM

 

Brian,

Lot more to be sure.

But we humans have become so complacent and reliant on our technology we often never notice.

We wake up to a micro chip beep, our coffee is brewed and ready for us, our TV is on and tuned to the traffic...and we are so just to it being that way we often don't know what to do when it fails.

 

We have become so dependent on our technology to keep us safe or to correct our mistakes, in spelling, grammar, art and math to name a few, that we have begun to lose those skills.

One of our most human traits, one that sets us above most of the other animals that inhabit this planet is the ability to choose, to make decisions based on intuition and experience, not blind instinct.

 

Which is why airlines have pilots and co-pilots, trains have engineers and conductors, and transit lines and trolleys have motormen...because the technology does fail, often because the person who programmed the computer failed to include every single scenario that could occur.

Reports have it that the motorman in the trailing put her train in emergency...one would assume because she saw the other train...or maybe she just had that hunch...we will never know because she died...but my point is, had the trains been under a different type of control, or a shared control, this might not have happened.

 

But my guess is that Metro, and the motorman running the trailing train, simply assumed that because the computer didn't take action nothing was wrong.

Complacency, being so used to the system always working right, will play a large role in this...but the fact she did plug the train while the computer failed to take action confirms that no matter how well thought out and programmed the computer is, there will always be a need for the human interaction and intervention from the cab.

And no, her putting the train into emergency did not prevent the accident, but I bet it saved a lot more lives than we will know.

23 17 46 11

  • Member since
    December 2007
  • From: Georgia USA SW of Atlanta
  • 11,919 posts
Posted by blue streak 1 on Tuesday, June 23, 2009 10:40 PM

blhanel

I suspect that we're going to find out that the computer lost track (no pun intended) of the train that was rear-ended.

   Computer control  --  OH Boy!!  First the obvious. Track detection circuit failed to note train that was run into.

2. Computer control  -- Why did computer not have software that will shut down area if a train gets lost?

3. Does the system operate in parallel with another computer and if not why? and if so why not shut down when there is a discrepancy?

4. Was a software patch installed sometime before this crash?

5. Three trains in a block? Did this fool the computer?

6. Computer power glich?

7. Is this putting too much reliance on one system without an independent backup?

8. Failure of an IC?

9. GIGO? (garbage in garbage out)

10. How many more possible computer problems can we come up with?  

  • Member since
    January 2002
  • From: Cedar Rapids, IA
  • 4,213 posts
Posted by blhanel on Tuesday, June 23, 2009 9:57 PM

I suspect that we're going to find out that the computer lost track (no pun intended) of the train that was rear-ended.

  • Member since
    July 2003
  • 964 posts
Posted by TH&B on Tuesday, June 23, 2009 9:27 PM
Can a computer train even travel at restricted speed ? (stopping at half the range of vision) Most trains in the history of railroading have been manualy controled, so most accidents have been due to manual operation. Perhaps in the future when most trains will be computer controled , most accidents will be computer accidents. This would not be the first serious train wreck caused by the computer running away with the train.
  • Member since
    January 2002
  • From: Canterlot
  • 9,575 posts
Posted by zugmann on Tuesday, June 23, 2009 9:15 PM
For this one computer fault (if it was the computer's fault), how many crews have hit other trains due to not obeying restricted speed? Exactly. .

It's been fun.  But it isn't much fun anymore.   Signing off for now. 


  

The opinions expressed here represent my own and not those of my employer, any other railroad, company, or person.t fun any

  • Member since
    April 2003
  • 305,205 posts
Posted by Anonymous on Tuesday, June 23, 2009 8:58 PM
 
henry6

Computers...they are employed because they are foolproof, don't take coffee and lunch breaks, don't take vacations, don't need health insurance, don't need social security and pension benefits, and don't pick up a pay check.  But are they properly fed and cared for?  .  I have been waiting for such a catastrophic meltdown to see what happens next.  Initially here, talk was to look at the train...now I see they are looking at the main computer back at the ranch.  I feel we are going into something big...bigger than texting and cell phone misuse.  Stay tuned.

 

henry6,

I think you are on to something.

Join our Community!

Our community is FREE to join. To participate you must either login or register for an account.

Search the Community

Newsletter Sign-Up

By signing up you may also receive occasional reader surveys and special offers from Trains magazine.Please view our privacy policy