Metro Crash: NTSB Investigation
Experts Suspect Failure Of Signal System, Operator Error

Peter Goelz
Fmr. Managing Director, National Transportation Safety Board
Tuesday, June 23, 2009 12:00 PM

Experts familiar with Metro's operations focused last night on a failure of the signal system and operator err as likely causes of yesterday's fatal Red Line crash.

Peter Goelz, former managing director at the National Transportation Safety Board, was online Tuesday, June 23, at Noon ET to discuss the investigation.

Metro was designed with a fail-safe computerized signal system that is supposed to prevent trains from colliding. The agency's trains are run by onboard computers that control speed and braking. Another electronic system detects the position of trains to maintain a safe distance between them. If they get too close, the computers automatically apply the brakes, stopping the trains.


Peter Goelz: Welcome to a discussion on yesterday tragic Metro Red Line train crash. Peter Goelz here to try and answer any questions you might have


Concerned in P.G. County: First things first: What the heck happened? Why did it happen? And what's to prevent it from happening again?

Peter Goelz: That's what the NTSB is going to determine. Their lead investigator is very experienced and a solid railroad man. I think there will be a lot of attention given to the automatic train control system and the operator's actions.


Washington, D.C.: I was talking with a friend from Boston last night who said that there was an MBTA crash up there a few months ago that involved an operator sending text messages on a cell phone that distracted him or her from operating the train which then crashed. The MBTA now apparently bans all cell phones from operators. I haven't heard what the operator error was in this incident, so I'm not saying a cell was even involved. BUT -- could the NTSB pass laws that ban cell phones from all operators at controls of trains, subways, etc? It seems perfectly logical to want operators to be operating the train and not phone, but I'm not sure of the authorities of NTSB in different jurisdictions.

Peter Goelz: Operator distraction- i.e. texting, phoning, iPods, etc- are a real problem in the transportation business. I am sure the investigators will examine whether distraction played a role in this accident.


Boston, Mass.: We've had big issues up here with MBTA train and bus drivers using cell phones and PDAs while driving including right before an accident. Any word on whether the Metro driver owns a cell phone/PDA, had it with her at the time of the accident and whether any messages or calls were made right before the accident? What is the Metro's policy on this?

Peter Goelz: I don't know what the specific Metro policy is, but I have to assume the use of any kind of device that could distract the operator is prohibited, particularly given the recent accidents where cell phone use played a major role.


Washington, D.C.: I know you are going to be flooded with comments this morning. But I think it is ridiculous that Metro and whatever other agencies responsible for information are playing the "ongoing investigation" card and withholding information regarding the cause of the crash. Also, why is the NTSB wasting their time posturing by blaming Metro for the lack of a particular safety improvements? This is not relevant, and it is extremely disingenuous by the NTSB to offer this up as its first finding. After the most recent train accident in Los Angeles, the authorities released information about the cause the day of the crash. In the present situation, Metro and the NTSB are sitting on information -- train speed, etc. -- that may point to the likely cause. They need to stop acting government bueracrats and be more forthcoming.

Peter Goelz: I disagree. The NTSB is one of the most public-access friendly agencies in the government. If they have factual, and I mean factual, information, they will release it.


McLean, Va.: From what I've read, the train's movements are fully automated, and it appears that the train operators have little to do during their shift besides opening and shutting the doors at the stations. I can see how this could lead to inattentiveness among all train operators. What can Metro do to raise the attention and awareness of its operators given the unstimulating nature of their job?

Peter Goelz: That's a good question. Operator fatigue and inattention is a problem that affects more and more areas of transportation as systems become more automated. There has been some studies on how to counteract boredom and inattention but more needs to be done.


McLean, Va.: How does Metro's safety performance compare with those of other major subway lines, such as those in New York, San Francisco, and London?

Peter Goelz: Many of these systems are showing their age and infrastructure funding to upgrade has been hard to find in many cities. The DC Metro record is not unblemished but is certainly not one of the worst.


Washington, D.C.: Is the D.C. Metro unique in its decaying infrastructure, or are most U.S. rail systems as broken (literally and figuratively)? If the D.C. Metro is a unique example, what are the steps to upgrade and increase safety measures? Where is the stimulus money?

Peter Goelz: One report indicated that there were upwards of 7 billion dollars of upgrades needed for the DC system. I don't know whether that number is correct but clearly more resources need to be directed to Metro. Simply retiring the older cars immediately would be a good step.


Fairfax, Va.: Locomotives have a safety feature to prevent one from riding over the top of another during a head on collision. Is there any way this device could be adapted for Metro railcars?

Peter Goelz: The crashworthiness of the cars will clearly be a major focus of the NTSB. They have great engineers and metalurgists.


Potomac, Md.: Thank you to the Post for excellent coverage of this tragedy. This morning's print edition states that "an eerily similar" incident occurred four years ago, but that "it was unclear last night" whether an investigation launched by Metro "ever found a cause." Has any more been learned about this incident and investigation? What are WMATA's procedures for making public the existence of such incidents and the results of investigations?

Peter Goelz: I am looking into that event now to see what lessond it may hold


Union Station, D.C.: What things will investigators do to compensate for the second train's lack of recording devices?

Peter Goelz: They will likey reccommend that these older cars be retired


Washington, D.C.: How long do such investigations usually take?

Peter Goelz: there will be a lot of activity over the next few weeks--the entire accident investigation may take 12-18 months


Rockville, Md.: Seems as though an attentive operator would have had to have seen the stopped train. I'm not suggesting that this is the cause, but I do have to wonder if WMATA has an official policy on the use of "instant messaging" devices by its on-duty employees. And, if so, what specific steps are taken to insure its observance?

Peter Goelz: Metro should ban all such devives


Washington, D.C.: From today's Post: "Debbie Hersman of the National Transportation Safety Board said the safety agency asked Metro to add data recorders after a 2004 crash and found it unacceptable that the system had not done so."

1) Does NTSB have authority to require such recorders; 2) Do you have any idea why Metro did not install the recorders; 3) Is it extremely expensive to install these recorders?

Thank you

Peter Goelz: The NTSB can only make recommendations --it does not have the authority to mandate


Takoma Park, Md.: I commute via the TP station every day. Today I drove to work. What can people who get on/off the TP station expect over the next few days, ie, shuttle buses, etc.? Thank you.

Peter Goelz: can not help you on this --i assume it will 2-3 days till they remove the wreakage at least.


Washington, D.C.: Apparently no "black boxes" were installed on the frontmost, stopped train. Can you outline the types of information gathered by those boxes and any other means of gathering that data in the absence of installed boxes?

Peter Goelz: the "event recorders" monitor only a linited number of actions--speed, brakes. Without the recorder they will do it the old-fashion way, measuring stopping distance and looking at the damage.


Rockville, Md.: Do the Metro trains have a "dead man" switch that would stop the train if the operator lets go of it?

Peter Goelz: I believe that they do when they are on manual operations. I do not know the system enough to speak to automatic operations.


Washington, D.C.: Apparently, NTSB made an earlier recommendation to Metro regarding either phasing out its older trains or equipping them with some kind of "black box" event recorder. As I understand it, Metro did not follow through with the recommendation. Could you please discuss this recommendation and what it means that Metro did not follow it. Thank you.

Peter Goelz: It means that Metro had other priorities for the expenditure of funds.


Washington, D.C.: I read earlier today that the train that caused the collision from the rear was given a "train is ahead of you message". Afterward the train proceeded its route. What could have happened?

Peter Goelz: I do not know if that is a fact, but I believe it will come out in the investigation.


Hyattsville, Md.: The NTSB report about the 2004 incident at Woodley Park faulted Metro for not having data recorders on the trains. What do you make of the fact that the colliding train in yesterday's accident appears not to have had such a recorder?

Peter Goelz: It means that Metro likely did not have the funds to promptly replace all of the older cars.


Washington, D.C.: Yesterday I was stuck in the tunnel on the Red Line caused by two trains, one that I was on and the one in front of us, having mechanical/electrical problems between Van Ness and Tenelytown from 4:10 to 5:00. Could this have been a sign of a system wide problem?

Also since this problem was on the same line headed in the same direction as the trains that collided for 50 minutes shoudn't all drivers on the Red Line have been aware of that problem? It has been my experience on Metro that a problem at one end of the line will still cause in-between station stops and starts all through the line. I don't mean to say anything negative about the poor deceased driver but I would think that she should have known that there were delays on the Red Line and would have been more vigilant about the possibility of a sudden stop.

Peter Goelz: I am sure that the NTSB will be looking into these problems during the course of the investigation.


Bethesda, Md.: Regarding texting, etc. I ride from Grosvenor to King Street 4 days a week and I routinely see drivers texting, talking on cell phones, reading the paper and talking to other Metro personel. I reported this once to Metro with specific info on time, train, etc., and only got a blow-off e-mail. This is a serious problem in Metro regardless of whether or not it contributed to yesterday's crash

Peter Goelz: You are exactly right. These kinds of distractions have been ignored by transit operators. Putting aside this accident, as we dont know what has happened, I am sure it will be looked at.


Union Station, D.C.: Just clarifying my question. What will investigators look at to figure out what happened if the second train didn't have its speed, etc., being recorded?

Peter Goelz: They will use whatever available data and eye-witness accounts they can get to estimate where the train was and what it was doing.


Bethesda, Md. : How likely is it that brake failure could be a cause? Is a total brake failure plausible?

Peter Goelz: Total brake failure is unlikely as there is redundancy built into the system.


Silver Spring, Md.: I am like a lot of the people in the area who are anxious for more information, however, I appreciate that they are trying to be careful to only share information when it's absolutely factual. It's dangerous to share anything other than exactly the facts at a time like this, and while I am frustrated and want answers sooner than later, I try to give them the benefit of the doubt. They did seem reluctant to provide facts that would fuel speculation -- even if they were just the facts... (especially about whether it is standard for the operators to be on manual during rush).

On another note, will Metro provide it's employees with any counseling or support of that kind?

Peter Goelz: The NTSB tries to stick to factual information only. Yes I believe that Metro will be providing counsuling services for their employees.


Arlington, Va.: Given the amount of wreckage in the first car, will NTSB be able to determine if the break system had failed to operate as a potential cause for the crash? Others have also mentioned older transit systems such as N.Y., London, and Boston and the need for upgrades and improvements. Given that Metro is relatively new compared with these older subway systems, is there a concern Metro is degrading at a faster rate?

Peter Goelz: The NTSB can learn a lot from the worst kinds of wreckage. The Metro system is over 33 years old and while it is younger than many other systems, it does need constant upgrading, and that costs money.


Operator Error: Does the NTSB have authority to request the phone records of train operators to investigate whether they were talking on the phone or texting at the time of the crash? While I realize this may not be the case in this instance I recall a relatively recent case in which a train operator was texting at the time of a fatal crash.

Peter Goelz: Yes. The NTSB has broad investigative power and they can secure this kind of information.


Rockville, Md.: You suggest that retiring older rail cars would be a "good step."

In reality, retiring 1/3 of the fleet would essentially mean that there would be major cuts in service, which would inevitably reduce ridership and put more people on roads. (It takes 5 years for new cars to be delivered once ordered, and there's no funding for new cars anyway)

In the end, from a statistical point of view, isn't it likely this would simply result in more road traffic, likely leading to more road fatalities?

Peter Goelz: That is a good point. I do think that Metro needs to commit to an absolute timetable to replace these older cars and that the city needs to commit the funding necessary to get the job done.


Rockville, Md.: Even if WMATA was unable, for economic reasons, to replace all of the old series 1000 cars, one might think they could have at least avoided using them as the first or last cars on any train, no?

Peter Goelz: I believe that the cars are operated in sets of two and they pair the older cars together as they pair the newer cars together.


Philadelphia, Pa.: Would a stainless steel car have fared any better in such a collision thean these aluminum cars?

Peter Goelz: The crash-worthiness of both trains, along with the emergency evacuation procedures, will be studied intently.


Downtown D.C.: You stated in a previous question that NTSB may recommend certain courses of action -- such as retiring older trains. Are those binding recommendations or optional?

Peter Goelz: The recommendations of the NTSB are recommendation period. However, about 85% of all NTSB recommendations are accepted and implemented.


Washington, D.C.: It's been reported that Metro initially indicated to the fire department that this was a minor incident and once upon the scene, the fire department immediately went into crisis mode.

Why would Metro under report a situation like this? Could they actually not have known the severity?

Peter Goelz: I have found that early information on any kind of accident is almost always wrong. The emergency response will be part of the accident report.


Metro Cars Construction: The last two crashes of Metro trains (Woodley Park and last night) involved cars telescoping onto each other. It appears from pictures of both accidents that the top of the car is wripped off from the bottom.

Are there different safety standards for the construction of these rail cars as compared to Amtrak Cars, for example, which appear to either break the impact with the vestibule area collapsing or jackknife (in other words, the passenger compartment tends to stay safer)?

In other words, is it safe to be riding in the front or rear car of a Metro Train?


Peter Goelz: That is a very good question. I know that there are different standards for inter-city rail cars, but I cant recall how the standards differ from transit cars.


Washington, D.C.: How many lives would have been saved yesterday if the older train had a "black box?

If the answer is none, why are we focussing so much on it when there's no reason to believe the cause of this accident won't be fully determined anyway?

Peter Goelz: The event recorder would only have helped us understand more clearly what had happened.


Philadelphia, Pa.: Is the manufacturer of the rail cars still in business?

Peter Goelz: I dont know. The newer cars were built by a Spanish company, I do not know about the orginal cars.


Alexandria, Va.: Recent reports indicate that the FAA routinely "balances" the financial considerations of airlines with safety recommendations. Now we have a transit agency who failed to follow even the most minimal changes to their oldest cars following NTSB recommendations a few years ago.

Do you think (Congressional) changes need to be made to give the NTSB some teeth to enforce its recommendations, or should that responsibility fall to some other agency? I fail to see the purpose of the NTSB carrying out very detailed, insightful, and presumably expensive investigations, if the results are ignored because implementing the necessary changes to too expensive or difficult.

Peter Goelz: That is a question that is often debated. I come down on the side that the present system works best, although I do believe that the media could give greater coverage to those recommendations that the NTSB makes that are not accepted. The NTSB needs a strong bully-pulpit.


Washington, D.C.: A black box wouldn't have prevented this accident. Prior to this accident a similar accident was investigated and a report issued. The best way to prevent this accident would have been to find the cause of the previosu similar accident and take appropriate actions. You say it's too early to know what happened here, but it sure isn't too early to know what happened in the previous similar accident. The report on the previous accident should be in the public domain now! Where is it?

Peter Goelz: The NTSB website and public inquires have all of the past NTSB reports, many of which are online.


Wheaton, Md.: Even with ""anti-crush"" frames, you still have 500,000 lbs at 50 MPH -- a lot of kinetic energy. What do the the "safer" designs do to get rid of that much energy.

Peter Goelz: Most recent designs involve locating structural collision posts that are designed to deflect the penetration of the car.


Kingstowne, Va.: Will it be possible to determine if the operator died as a result of the crash or if something incapacitated her prior (seizure, heart attack), so that she couldn't do anything about the impending crash?

Peter Goelz: Yes. They will do a full forensic examination.


Washington, D.C.: Do you think it would be a good idea if Metro adds on a sufficient number of cars in the upcoming 7000 series bids? Sounds like the 1000 series cars days are numbered.

Peter Goelz: Yes. The 1000 series cars should be retired as soon as possible.


Falls Church, Va.: There's been far too much negativity in this chat. The fact is, Metro is managed by the 2009 Public Transportation Manager of the Year, John Catoe. The American Public Transportation Association honors Catoe at their October meeting. I do hope people look past the mechanical difficulties from yesterday, and focus on all the wonderful work that Metro has been doing over these past three years.

Peter Goelz: Metro has a reasonably good record and the system has tried to function under very difficult budget constraints.


The report on the previous accident should be in the public domain now! Where is it?: Collision Between Two Washington Metropolitan AreaTransit Authority Trains at the Woodley Park-Zoo/Adams Morgan Station in Washington, D.C.November 3, 2004

Peter Goelz: check the NTSB data base on line


Bowie, Md.: Just out of curiosity, who thought that making humans the "failsafe" devices for automated machines was a good idea? I mean people make mistakes so why not reverse this and have the machines be the failsafes for the humans. The train operators would feel needed and do their job (pay attention) while the machines would be able to catch any mistakes that may occur. Automation is fine for some things, but we may go too far.

Peter Goelz: great question and it is one for our times


Fatalities: Were they all people who were in the first car of the train that went on top of the other train?

Peter Goelz: I ahve not seen a map of where they were found.


Washington, D.C. : What percentage of the fares collected by the Metro actually go back into the Metro? Is it possible that in the current economic downturn that more of the Metro money is being reallocated for other government programs?

Peter Goelz: I do not know what the split is-oe if there is a split


An Urban Planner in Maryland: Will the federal transportation bill that is working its way through Congress provide additional funding to transit agencies to upgrade their infrastructure (in the hopes of avoiding future accidents.) I know that Metro is blamed for not implementing your recommended safety precautions, but I also know that they don't. have. any. money. In such instances, do you recommend that they cut service?

Peter Goelz: that is a tough political question--sysytems are forced to operate under less than ideal circumstances-bur i do think cuts in service might force the hands of those with prse strings.

_______________________ This concludes today's discussion with Peter Goelz. Thanks for joining.


Editor's Note: moderators retain editorial control over Discussions and choose the most relevant questions for guests and hosts; guests and hosts can decline to answer questions. is not responsible for any content posted by third parties.

View all comments that have been posted about this article.

© 2009 Washingtonpost.Newsweek Interactive