Metro: Train Control System Malfunctioning
The train control system designed to prevent Metro crashes is malfunctioning across the railroad system, suggesting that a technological failure at the heart of last month's fatal crash might be widespread, according to officials and documents.
Washington Post reporters Lena H. Sun and Lyndsey Layton were online Wednesday, July 22, at Noon ET to discuss the findings and what Metro is doing about the malfunctions.
At least a half-dozen track circuits on four of the transit system's five lines have failed to properly detect the presence of trains in recent weeks, records show. The safe operation of a transit system requires that the location of trains be known at all times.
Lena H. Sun: Hello everyone. Thanks for chatting with us today. Hope your Metro rides today were relatively smooth. Let's get to your questions.
Hyattsville, Md.: How often does Catoe ride Metro? It's fine for him to talk about how safe it is when he never rides on it.
Lena H. Sun: Hello Hyattsville. Actually, Catoe rides the subway and bus regularly. Some of the Metro board members also ride the system regularly, but others do not.
Silver Spring, Md.: Count my wife and me as two more who have given up on commuting via Metro. We will ride when we have time on our hands, but not when we need to arrive somewhere on schedule. It is not a safety issue for us, but a reliability and quality of service issue.
Lyndsey Layton: Hi Silver Spring,
I think your frustration is shared by many, especially if you board in and around Silver Spring.
Alexandria, Va.: I understand that all trains are being operated manually. Other than a slightly less comfortable ride due to uneven breaking action when approaching station stops, what are some of the other disadvantages of manual operation?
Lyndsey Layton: Hi Alexandria,
You're right that manual operation causes a bumpy ride. But there's another downside. The Metro system was designed to run in automatic mode and when it is shifted to manual, you lose efficiency - i.e., trains cannot run as close together as the design calls for. That means fewer trains run across each line per hour, which means more crowding on each train and a slower trip.
Arlington, Va.: Catoe was on TV this morning claiming that this report is a big exaggeration. I assume the disconnection of these track circuits is not an on-going thing, is it? Trains that I have been on do not seem to be stopping for a long period of time in the Court House area. Have they been able to replace the malfunctioning sensors? Any idea why they have not been able to fix the circuits near the crash site? Can't they just replace the whole thing in that area? Certainly it would seem like something that is a worth a try to alleviate the current situation. I guess we are all mostly captive to the system. There aren't a lot of options, so we all just have to suffer along with the delays and problems. Metro talks a good game about improving communication with passengers, but they have been saying the same things for years, and nothing ever really changes.
Lena H. Sun: Hi Arlington. We heard that some track circuits were being turned off in recent days and we followed up with Metro. Metro's rail chief Dave Kubicek told us they found problems with some track circuits as part of this ongoing computerized test they are now running twice a day. In some cases, they were able to make adjustments and fix the circuits. But in other instances, where fixes could not be made immediately, they have disabled the circuits.
He did not know what is causing the problems or how long the circuits will be turned off. It is possible that in some cases, the issues could be related to track maintenance in the area.
Silver Spring, Md.: I don't understand the disparity between the information that came out in the days and weeks before and after the crash, and now.
Is this a case of Metro hiding information, the standards of what constitutes a failure/malfunction changing, or a case of 5 other circuits magically failing only this week?
The former, a cover-up, would seem more serious than if people are simply looking at things with a tighter tolerance now...
Lyndsey Layton: Hi Silver Spring,
Those are excellent questions. Our reporting shows that since the June 22 crash, Metro has been increasingly scrutinizing the track circuitry. Before the crash, the circuit data was checked monthly. After the crash, officials began examining it daily. And then about 10 days ago, Metro decided to look at the data twice a day - after each rush period. They also appeared to have lowered the tolerance for "erratic" behavior - flagging any circuit that is not performing correctly for longer than one second.
As we reported today, that has resulted in Metro flagging a half dozen circuits as problematic. In addition, we obtained records that showed another handful of circuits behaving the same way.
Meanwhile, Metro General Manager John Catoe has been saying publicly that the agency physically inspected the circuitry and found no other "anomalies' in addition to the circuitry at the crash site.
Silver Spring to Rosslyn Commuter: Is there any end in sight? The Metro was just atrocious yesterday, so I'm back to taking the bus down 16th Street. I'm almost tempted to take my car and brave the traffic in D.C., but I fear that will be worse. What do you think the final outcome of this "investigation" will be? Will Metrorail have to shut down since it will be deemed unsafe?
Lena H. Sun: It's tough for folks on your end of the Red Line. Investigators haven't been able to figure out why the track circuit at the crash site is still not functioning properly, even though they have replaced components. Until they do, Metro is going to continue running trains one at a time through that section of track between Takoma and Fort Totten, which causes delays up and down the line.
This is also the time of year when ridership is high because of visitors and tourists, so lots of people riding the train. Some people have found the new express S9 express bus to be a good alternative.
Washington, D.C.: From the article, has Metro explained this discrepancy?
The documents show that Metro technicians have detected malfunctions since at least July 11. Metro General Manager John B. Catoe Jr. said publicly as recently as Thursday that the agency physically inspected all 3,000 circuits and did not note any problems.
Was Mr. Catoe untruthful or under-informed?
Lyndsey Layton: When we were doing the reporting for this story, we asked Metro for some comment from Mr. Catoe. The agency did not supply one. He has yet to address this issue.
Vienna, Va.: Given the current issues with the trains, how slow do you expect rush hour to be tonight?
Lena H. Sun: Metro has been advising riders to add extra time. I've been hearing from riders that Red Line in the evening has been tough. Hard to say, because it all depends on which line you take and when you're heading home.
Lena H. Sun: Just wanted to chime in here and say that actually I did hear from Mr. Catoe, through spokeswoman Lisa Farbstein, when we were working on the story. He made some of the same points that had already been told to us by rail chief Dave Kubicek.
After our story appeared on our Web site, Metro issued a statement that basically made many of the same points officials had made before, and we did quote from that statement in our story.
Washington, D.C.: Do you know anything about the results of the sight-line testing at the accident site? Were the brakes applied as soon as it was possible for the operator to see the stopped train up ahead?
Lena H. Sun: The NTSB has not yet said anything publicly about the results of the sight distance test. I expect they will address this question at some point because many people want to know that answer.
Monitoring the trains in the system: I was hoping to get an idea of how the computer system showing all the trains currently running is monitored. In theory each train in the is visible, right (Which is part of why turning off the computer is so extreme)? How reasonable is it that a person, monitoring what's happening on the computers, could see that a train was not slowing down and in danger of rear-ending another?
I guess I'd like to know if someone is monitoring such a small area of the system that they could pick up on this and let the drivers both know or if they're responsible for great swathes of the Metro system so that it would be unlikely.
Lyndsey Layton: Hi there,
A crew of train controllers work in Metro's downtown operations control center and each controller is responsible for monitoring train movement on one portion of the 106-mile railroad. The controllers watch the train movements on a huge screen that is a schematic version of the railroad.
Metro officials have said that when the track circuit at Fort Totten malfunctioned and did not detect the presence of a train that was idling outside the station, it would have shown up on the downtown schematic map as a flickering light, but that it was flickering so quickly, the controller would have to be staring at it in order to catch it.
It's interesting to note that the track circuit in question had been flickering for five days before the crash.
Washington, D.C.: Are we looking at manual operation indefinitely? I mean, how long can we (the D.C. metro area) manage with this type of service? I have stopped using the metro, but find even the bus system is now suffering as a result. It gets to the point where I figure, I'd might as well just drive and add to the road-traffic. I hope that Metro officials and D.C. officials are up day and night trying to rectify this situation, urgently.
Lyndsey Layton: Hi Washington,
We hear you. And as a rider of the S line on 16th Street, I can tell you that those buses headed for Silver Spring have been packed. We wish we had an answer but haven't heard anything from Metro officials on how long this is expected to last.
Circuit data checked only monthly before the crash: I'm not an engineer at all, but is length of times between tests really reasonable? This is an integral part to the safety of the system, and it was checked only once a month? It makes me very nervous that presumably the law allowed Metro to test only once a month ...
Lyndsey Layton: Hi,
Actually, there is no law that spells out how frequently a transit system must test equipment, or how it should be maintained, or basically how a transit system should operate. Transit systems in the U.S. are largely unregulated.
From Dr. Gridlock: Hi, Lyndsey and Lena, if a track circuit is having a problem detecting the presence of a train, does it make any difference whether the nearby trains are under automatic control or manual control? Is one type of control more likely than the other to avoid a crash?
Lyndsey Layton: Hi Dr G! It's not supposed to make a difference. The track circuitry that keeps trains from colliding is a separate system from the computerized controls that run the trains.
Washington, D.C.: "He did not know what is causing the problems or how long the circuits will be turned off. It is possible that in some cases, the issues could be related to track maintenance in the area."
Can you explain in more detail what the last statement means -- how does track maintenance adversely affect the detection circuits? Thank you.
Lena H. Sun: I wish I had more detail. Unfortunately, Mr. Kubicek did not provide much elaboration, except to suggest that perhaps maintenance work might have done something to connections. But that was only one possibility. He did not have an answer.
Beautiful Silver Spring, Md.: So how scared should we be -- of another crash, or alternatively of service disruptions that send people into their cars and throw the roads into further paralysis?
Without Metro, this city doesn't work.
Lyndsey Layton: Hi Beautiful SS,
I think Metro is probably operating at its safest right now - they've flagged the circuits that are not operating properly and only allowing one train at a time through those sections at reduced speeds. The only thing troubling about that from a safety perspective would be the fact that when you disable a track circuit, you also disable the system's ability to detect a break in the rails. But we know that Metro crews are inspecting those areas thoroughly anyway now, trying to solve the circuitry problem.
The issue is, how long can you sustain this, without doing as you suggest - sending riders onto the roadways.
Washington, D.C.: With all the trains running on manual and all the extra braking (stop and go) and wear and tear that that involves, what are the additional maintenance costs? And is Metro monitoring the safety implications of this?
Lena H. Sun: I don't know what the extra costs are but that is certainly going to be a factor they need to monitor. Some safety experts also point out that manual operation, while it keeps operators more alert, does introduce more hazard into the system. There is a reason why train systems are optimally automated, they say: human error causes the vast majority of incidents and accidents.
Alexandria, Va.: I have notice that some trains are still running the older cars as the first and last car. I was under the impression older cars would be moved toward the middle. Is this not the case?
Lena H. Sun: Metro has said they would try to put the older cars in the middle as much as logistically possible, but it may not always work out that way.
Silver Spring, Md.: As a Red Line traveler from Silver Spring to Dupont, I'd be curious to know whether I can expect the Red Line to be back to normal after Labor Day. I am not looking forward to the persistence of longer travel time, unpredictable train timing, and substantially more crowded trains.
Lena H. Sun: Hi Silver Spring. It is really hard to say. It all depends on the investigation. When Metro had a problem with relays some years back, trains ran in manual for nearly two years.
Red Line: It doesn't matter WHICH end of the red line you are on, there are delays every morning and every night with impossible crowding! I commend Metro on one thing - yesterday during the morning rush hour there was a delay at Medical Center area - instead of sending jam packed trains through to the rest of the stations downtown like they usually do, Metro sent an unoccupied train to the Cleveland Park station which gave those who live in the District a chance to board the train just like everyone else. So I thank them on that.
Lena H. Sun: Okay, thanks for your comment Red Line.
Day-After Post Story: I think it was the two of you in the by-line of the article that was splashed all over the landing page the morning after the accident. The headline read that operator-error was an area of focus, and the content of the article blindly quoted an "expert" who said she (the train operator) should have been able to see the train and stop it in time.
That article has been bothering me since I saw and read it. First, I think it's obvious that the equipment and the operator would be focused on in any investigation of any crash involving both. Second, she has so far not been implicated in any way that I have read. Yet the morning after she suffered a pretty grim death, the Post offered unattributed opinions of her contributions. I wish you hadn't.
Lyndsey Layton: Hi,
Thank you for your question. Those first stories were written about four hours after the crash, when we were working as hard as we could to make sense of the chaos. You're right that the equipment and the operator are the two immediate targets of any crash investigation. From the impact, it was clear that the striking train was traveling at some significant speed as it hit the idling train. We quoted a rail safety expert who said "It doesn't look like she hit the brakes."
We did not identify that expert because that person insisted s/he could not be identified since the crash was under federal investigation.
That's a difficult balance to strike. Readers want to understand why the accident happened and we try our best to try to explain the thinking of the best informed sources we can find.
Two days later, when the investigators told us that the emergency brake had been deployed, we ran a front page story to say the operator had applied the brakes.
We still do not know if she applied them in a timely way, or if she hit them too late to avoid the accident. We are waiting for the NTSB to disclose the results of sight distance tests that will answer that question.
Adams Morgan: Catoe is on WAMU right now talking about how the Post is distorting things. Who is doing the distorting? Who is the public to believe?
Maybe you guys can call in and ask him why he kept talking about no problems when there were problems? Was he aware of the problems or did his staff not inform him? Who is being held accountable?
Lena H. Sun: Hi Adams Morgan. Our story reported that Metro has been conducting twice daily tests to look for problems in its track circuits. This is a higher level of scrutiny that began after the crash. As a result of the review, officials flagged track circuits that were not performing correctly for more than one second. In some cases, they made adjustments and fixed them, according to Metro's own rail chief, Dave Kubicek.
In other cases, according Kubicek, they turned off or disabled the circuits. In those areas where the circuits are turned off, trains have to run at slower speeds.
Anonymous: Hi, I've heard metro claim that the technology that they need for the trains does not yet exist. However, I've been told by people in the human-centered engineering field that the technology does, in fact, exist, specifically in the military (no, it's not classified).
How much of this is the politics of metro and it's acquisitions/contracts? Specifically, has metro in the past or currently taken bids by large AND small businesses who may actually have developed this technology?
Bottom line: does metro know how to buy the things it needs--specifically high tech stuff?
Lyndsey Layton: Hi Anonymous,
Metro is technically correct when it says that there is no commercially available system it can go out and buy and install here. That's because no two transit systems are alike - they all run under different operating conditions and have different kinds of trains. So anything Metro buys has to be tailor-made. That doesn't mean the technology doesn't exist - it does. It just has to be made to fit the Metro system. Which is a lot easier than having to invent something from whole cloth.
Washington, D.C.: I have read that up to 50 percent of all rush hour commuters work for the Federal government. Given that, it would appear that the Federal government has a huge stake in Metro operating safely and smoothly. But I haven't heard of any concerns expressed by Congress or the White House or OPM. Have you? Or is this because all the political hotshots haven't taken Metro every in their lives?
Lena H. Sun: Yes, close to 40 percent of all rush hour riders are federal workers. Folks in Congress, especially the regional delegation, have been pushing hard to get a steady stream of funding for Metro because of its unique role here. In the past, these efforts have been blocked by some key senators. Unclear if those senators have ridden Metro. But I would guess many of their staff members do.
Lena H. Sun: Hey folks. Thanks for all your questions and comments. That's all we have time for today.
Editor's Note: washingtonpost.com moderators retain editorial control over Discussions and choose the most relevant questions for guests and hosts; guests and hosts can decline to answer questions. washingtonpost.com is not responsible for any content posted by third parties.