Transit
The Price of Safety, part 2: Recent derailments expose problems
Previously, we looked at a rash of accidents involving Metro track workers. A common theme across many of those incidents was a problem with the implementation of safety rules within Metro.
Aftermath of Mt. Vernon Sq. derailment. Photo by KeonT.
Since 2003, Metro has experienced a significant number of derailments. Many of these incidents involved out-of-service trains in rail yards, or on segments of pocket tracks. There are several notable incidents in recent history that have involved in-service trains, including one that resulted in passenger injuries. The causes of these derailments vary, though one factor remains the same--a failure to implement safety improvements throughout the system.
On January 20, 2003 the last car of a Blue Line train derailed along the elevated track outside of the National Airport station. Forty-six passengers were safety evacuated on that frigid night, though the damage totaled over $100,000. Over the next five years, there would be at least nine more derailments. In June of 2005, amid the rash of derailments, the Washington Post published an investigative piece entitled "Safety Warnings Often Ignored at Metro." The feature piece pours over internal Metro documents, as well as accident reports, and comes to some startling conclusions. Many of these were eerily prescient. After the National Airport crash, an internal investigation determined that employees were aware of the potential for problems. Documents showed that track managers expressed concerns relating to the type and condition of the track being used at the location. According to interviews, those concerns were passed up the chain to upper management, though no action was taken.
Warnings about other causes of derailments, such as a lack of track lubrication were also ignored. Investigators determined a lack of lubrication caused a derailment at the Alexandria rail yard. There was a brief push towards getting track properly lubricated, but over time the efforts began to falter. Don Painter, former manager of the track department, told the Post that he blamed a lack of institutional memory. "The superintendent retired, the assistant superintendent went to a different location, the maintenance manager went somewhere else, and the guys, when no one told them they needed to keep lubricating, the ball got dropped."
It took a year and several derailments later for Metro to implement directives regarding track lubrication.
Metro's track department was rife with systemic problems, especially concerning track inspections. The Post's investigation turned up a 2004 audit that showed some track workers did not know how to report problems to their superiors, and that inspectors often overlooked problems because they were required to cover an unrealistic amount of track each day. Additionally, prior to 1999 there was no official training program for track walkers, and audits showed that even after programs were implemented there was a lack of proper training.
In 2005, Susan Coughlin, a former NTSB member, noted that these incidents were "indicative of systematic oversight problems which, if left unaddressed, could produce a catastrophic accident."
Nearly four years after the Blue Line derailment, and 18 months after the Post article, a Green Line train derailed outside of the Mt. Vernon Square station. At 3:45 pm on January 7, 2007, the fifth car of the six-car train derailed, injuring 23 people. The NTSB determined that a problem with wheel maintenance caused the derailment. In the Railroad Accident Report, the NTSB also identified systemic problems within Metrorail. From the the report, RAR-07-03,
WMATA was aware of the wheel climb derailment problem with the 5000-series cars before this accident. Transit industry research and discussions with WMATA management indicate WMATA was aware of work done by the Transportation Research Board for the National Academy of Science and National Academy of Engineering on flange climb derailments in transit operations. Additionally, WMATA commissioned, participated in, and received the final Wheel-Rail Interface Study from the TTCI. Extensive testing to determine the cause of these relatively similar derailments in the 5000-series cars failed to produce a solid answer. Also, the APTA panel concluded that there was no single cause in the seven derailments it examined; however, the panel did identify several specific factors and made recommendations for WMATA to consider to prevent future derailments. Nonetheless, after requesting reviews by industry experts and funding related research work, WMATA failed to effectively address the proposed safety recommendations before this accident. The Safety Board therefore believes that WMATA should establish a process, including a single point of responsibility, to prompt timely evaluation and action on proposed safety improvements that are identified as a result of accident and derailment investigations and related research projects.In June 2008, an Orange Line train derailed outside of the Courthouse Metro station. There were no injuries, but over 400 passengers needed to be rescued from the tunnel. Significant damage was done to the track in the area. Interestingly, the train operator did not immediately realize the derailment had occurred. A Metro supervisor happened to be riding on the train, felt a jolt, and alerted the train operator. The train had traveled more than 2,300 feet with the front wheels of the third car off the track.
An internal Metro investigation determined that a track walker had failed to report a significant problem with the track in the area. From a WMATA statement:
"Our inspector failed to recognize the out-of-tolerance rail conditions," according to Metro's Chief Safety Officer Ronald Keele. "The track conditions compounded with the forces of the moving train caused one wheel to climb atop one track and the other wheel to drop to the ground. We are very fortunate that there were no injuries."The inspector was suspended for five days as it was his first offense. Again, Metro stated they would work to revamp their training for track inspectors.The findings of Metro's internal investigation determined that the track inspector failed to detect "defects in the track's geometry" in the area of the derailment and "violated several inspection procedures," according to Keele.
The inspector did not inform the Operations Control Center of dangerous track conditions, did not report any dangerous or defective conditions to his supervisor, and failed to properly measure the space between the two tracks.
Metro's safety record regarding derailments highlights significant organizational deficiencies. Time and again, there was prior warning that problems could develop. Time and again, these warnings went unheeded. The same pattern emerges when looking at prior collisions and "near-misses" on Metro. Again, we see warnings from both within Metro and the NTSB that were ignored. It was no secret that trouble was brewing.
Dave Stroup is an online organizer and progressive activist. He enjoys public transit, Democratic politics, and rabble-rousing.   |
Comments
- Bikeshare is a gateway to private biking, not competition
- Short-term Washingtonians deserve a voice, too
- Judge denies injunction against closing schools
- DC Council makes major policy changes overnight
- Public land deals have both benefits and pitfalls
- Long-term closures: A solution to single-tracking?
- PG planners propose bold new smart growth future
The themes keep repeating themselves: previous warnings, generally unheeded, no real accountability, heads never roll, and the same tragic errors and consequences continue.
Perhaps Congress - which created the framework for Metro - should change the law to place Metro safety issues under the supervision of a federal agency. If they lack the necessary culture of safety, maybe someone from NTSB should be Metro's Safety Czar. (After all, we need all the czars we can get in this town.)
The DC, MD, and VA members of the board seem more involved in turf battles and jurisdictional politics. An outside safety expert - with real powers - may be what's needed.
Again, solid work.
by Mike Silverstein on Oct 14, 2009 10:21 am • link • report
I saw some track walkers yesterday. It's pretty alarming to see the tunnel get bright behind them, even when you see them step out of the way.
by цarьchitect on Oct 14, 2009 10:24 am • link • report
by Davin Peterson on Oct 14, 2009 11:08 am • link • report
As for cars that are "too old", most of the systems that opened before Metro run some cars of an earlier vintage than Metro. The original BART cars, the PA1-PA3's at PATH, the R32/R40/R42/R44/R46 in NYC, the 01500/01600's at MBTA, and the 2200's/2400's at CTA, the H4's/H5's at TTC, and the entire fleet of the STM are the same age or older than the 1000's. Fatal accident or not, I don't hear much of a hue and cry to get them out the door and some of those cars don't have definite replacements on order.
by Jason on Oct 14, 2009 12:08 pm • link • report
Proper understanding of the human factor in operating and maintaining the railroads is crucial in improving the safety at Metro. Often times, the immediate supervisor is tasked with getting his subordinates trained, but may be limited to an inadequately staffed training department.
The age of the 1000 series cars is a concern, but at this point, only in terms of reliability. If you do a bit more research, you should see the newer subway cars with the rigid frame and bogies have much more to do with the derailments.
by retired metro manager on Oct 14, 2009 12:14 pm • link • report
by Dave on Oct 14, 2009 12:23 pm • link • report
Perhaps a slow moving train w/ cameras and lights on the front that allows someone in the cab and then again at a remote location to review the track.
And why is so much of the knowledge in the heads of workers?
Write it down! ISO 9000 is good for something.
by shy on Oct 14, 2009 12:40 pm • link • report
by Dave Stroup on Oct 14, 2009 1:12 pm • link • report
by Dharm on Oct 14, 2009 1:28 pm • link • report
There's no reason track inspection can't be accomplished within the current hours available.
by Alex B. on Oct 14, 2009 1:33 pm • link • report
by SJE on Oct 14, 2009 2:08 pm • link • report
The problem is that Metro's Board is too political and doesn't want to make any difficult decisions. In Tokyo, professionals not politicians run the transport system. I know I was there and saw girls in high heels run like the wind on a Friday night to beat the midnight closing time (or face a VERY expensive cab ride, you step in a cab in Tokyo and it's like $5).
by Dharm on Oct 14, 2009 2:39 pm • link • report
Starting with the basics are the track gauges, the dimensions of the subway tunnels and platforms, and the supply voltages. There are also other factors with the current infrastructure and what can be supported, such as the type of train to wayside signals for Automatic Train Protection (ATP) and Control (ATC).
Another issue is whether or not there is concern for having a much heavier car with stainless steel or a lighter aluminum car. It is usually not a good ideal to mix the two for safety reasons, unless you can appropriate them to separate lines to minimize collision issues. You really don't want a stainless steel car running into an aluminum one.
Some other concerns might be that BART, for example, can accommidate a 10-car train consist into a station, while WMATA can only accommodate an 8-car consist. Also, I believe the BART cars are a bit wider than the WMATA cars.
Depending on how your maintenance shops are set up, it might be a factor on what kind of equipment is installed, and what kind of access is required. Also, there is a concern regarding what the capability of the workforce is for repairs, other than routine repairs. Maintenance contracts might be needed for one agency and not for the other.
There are ongoing endeavors to combine the needs of agencies to overcome these issues, and one case involves the new 7000 cars for WMATA. At one point in time, there was going to be a combined purchase with Miami. However, the operating console actually faced the opposite side for the Operator, so there would have to be some strategy to overcome that. In the end, the Miami Board did not approve the combined purchase. It would have been interesting though, since there would have been savings in a lot of the engineering work.
Hope that helps some.
by Retired Metro Manager on Oct 14, 2009 3:06 pm • link • report
The daily inspections by the track walkers are needed to spot changes in the track conditions that might occur due to changes in weather, obstructions/debris on or near the tracks, damaged or missing 3rd rail covers, or other unusual conditions that can not be spotted from the operating train(such as in curves).
I know I'm missing other things but thought I'd at least get a few notes in on this subject.
by Retired Metro Manager on Oct 14, 2009 3:17 pm • link • report
NTSB dodged the underlying issue in the Mt. Vernon derailment. The 5000 cars were designed to be lighter in order to use less electricity. However, when actually employed on the tracks, the reduction in weight caused the cars to be top heavy. When the cars moved around curves, this caused wheels to climb.
WMATA debated how to deal with the issue for several years including considering hanging sand bags on the undercarriage of the cars to better balance them. Nearly all derailments have been 5000 series cars in the past several years.
After the Mt. Vernon derailment, NTSB recommended "guarded turnouts" as a solution. This is a mechanical solution that keeps the wheels from climbing off the rail at crossovers or curved sections of track. WMATA has been slowly replacing all unguarded turnouts with "guarded turnouts" but there's plenty of curved track that is still unguarded.
On another note, I'm told cab radios installed in the newer Metro cars still don't work properly. Operators have to rely on hand-held, fixed phones along the right of way or when outside (gasp!) sometimes cell phones.
by kreeggo on Oct 14, 2009 3:21 pm • link • report
by Retired Metro Manager on Oct 14, 2009 3:24 pm • link • report
by Retired Metro Manager on Oct 14, 2009 3:29 pm • link • report
Thanks.
by Dave Stroup on Oct 14, 2009 3:47 pm • link • report
Without safety, all our dreams about expansion are futile efforts in fantasy.
Institutional memory: this is one problem with DC being so transient. But still, for all the people here with multiple degrees, many times business or management degrees, you'd think that basic principles would be followed regarding preserving knowledge, and as someone else pointed out ISO 9000.
Please - more like this one GGW!
by Jazzy on Oct 14, 2009 7:15 pm • link • report
by Turnip on Oct 14, 2009 7:36 pm • link • report
by kreeggo on Oct 14, 2009 7:42 pm • link • report
Union Station opened in 1908. I assume you're referring to its centennial celebration from last year. Even Washington's not 300 yet. In fact, modern rail transport only dates back to the 1820s.
by Matt Johson on Oct 14, 2009 10:24 pm • link • report
As an aside, in the proceeding sentence Mr. Catoe remarked that the switches that were replace in the Arlington Cemetery interlocking in January of 2007 date back to 1977. The information given to the Mr. Catoe about the age of the switches is incorrect. I took photographics of the turnouts, on the weekend they were removed, one photos shows a 09 1993 mill mark on one of the turnout stock rails.
by Sand Box John on Oct 14, 2009 11:55 pm • link • report
As a retired Systems Maintenance(SMNT) technician who spent over 27 years at Metro, I know from bitter personal experience that change comes slowly (if ever) to Metro. It is essentially a socialist organization. Everyone gets paid about the same amount of money +/- maybe 15% (with a few exceptions) regardless of what they do or how hard they work. Incompetent decisions that cost Metro millions of dollars go unpunished. Very few people care because it's not their money -- it doesn't come out of their paycheck. There is no 'carrot', and often, not much of a 'stick' either.
In addition, Metro answers to no one except their own board. They don't need no stinkin' badges. Oversight is all but nonexistent.
Also, Metro's safety department has the same insular culture as the rest of the organization. For example, I heard recently from a former coworker that he emailed the head of Safety, Alexa Dupigny-Samuels, and was rebuffed by one of her underlings. His email was intended to be a friendly hello and was treated as if it were hostile. I was once screamed at by a safety dept. employee for having the audacity to report safety violations that her dept. had specifically asked us to report! Most of their talk about safety is just that -- talk.
Their response to the recent rash of employees getting hit by trains was to issue hard hats and ANSI 'class 3' safety vests. The hard hats are pointless (unless working around cranes and other overhead hazards) and the vests are so large and loose that they practically fall off of many employees and the loose half sleeves are easily snagged on wayside equipment, and possibly, passing trains. In addition, the Velcro that's used is so strong that these vests do not easily break away. They are _more_ dangerous than the old 'class 2' vests. They actually violate Metro's own safety rules regarding loose fitting clothing!
It's all window dressing.
For those reasons (and others) I would say that no one should hold their breath waiting for improvements in safety at Metro.
BTW, the only reason to lubricate rails is to prevent wear, not derailments. Some of the train yards have automatic rail lubrication on turns with a tight radius, but lubrication should never be necessary to prevent derailments. If it is, there is something fundamentally wrong with the track layout. Also, consider that much of the lubrication that Metro management decided was necessary was being done periodically, by employees. If it _is_ critical, then there are a few problems with that plan:
1) Different people may not apply grease the same way and/or to the same areas of the rail.
2) In above ground locations, the weather has a huge affect on how long the grease remains on the rails.
3) The number of trains running over a particular stretch of rail will also affect the longevity of any lubrication.
It is unrealistic and unsafe to rely on grease (particularly manually applied grease) to prevent train derailments.
To my knowledge, there are no other transit systems or railroads that use grease to attempt to prevent derailments.
Sadly, the comments about Metro management ignoring recommendations and concerns -- both from outside agencies like the NTSB and from employees -- are all too true. That attitude continues to this day, and I don't see any way to change it, short of firing (or severely punishing) those responsible.
I enjoyed reading both of your posts in this series and look forward to the next one.
by Retired SMNT employee on Oct 15, 2009 4:43 am • link • report
The issue is not age it is safety. WMATA knows, and has known for years that the 1000-series cars are not crash-worthy and pose an undue risk people riding in them.
When a model of car is found not to meet crash safety standards a recall is ordered and the manufacturer is forced to repair or replace it regardless of the negative effect on the finances of the manufacturer. When a safety problem is found with a model of airliner the FAA can (and does) order it grounded until fixed regardless of how this will effect the company that made it or the airlines that operate it. I fail to see why Metro should be allowed to continue operating equipment it (and the NTSB) knows is not safe for years merely because replacing it is difficult and expensive.
Earlier this week, Ford recalled over 14 million vehicles over a safety issue. I suspect that if they, instead, had announced "We've known about this problem for several years but the economic downturn really hurt our bottom line so we are going to delay the recall until 2012" the response would have been an NTSB mandated recall and class-action lawsuits from their customers. Why shouldn't Metro be held to the same standard? (and no, I don't consider they don't have enough money to be an acceptable answer)
by Jacob on Oct 15, 2009 1:07 pm • link • report
WMATA has utilized this same "recall" philosophy to it's railcar fleet. Although rare, fleets have been grounded. If not grounded, many conditions can be isolated within the car, but with restricted operating parameters in place. For example, there has to be a large percentage of braking (over 85%) for the consist and over 50% propulsion. Do we really want to ground 35% of the WMATA fleet at one time?
All the current rail fleet of cars do poorly in collisions. They will climb on each other and other obstrutions, such as the bumping posts in the yard, at fairly low speeds. This has been demonstrated with past collisions with like cars, not involving the 1000 series fleet. The 1996 collision at Shady Grove demonstated that, when a six-car train of Breda cars collided with a four-car gap train of Breda cars.
The NTSB pointed out the 1000 series cars have less collision protection than the other fleet of cars, but this is really to a small degree. You will get one car climbing over another and the car body telescoping with collisions at low, moderate, and high speeds.
WMATA has a lease agreement with some major banks to keep the Rohr cars (the 1000 series) up thru 2014. The real question may be what is the useful life of a rail car and when does it become a reliability issue, and of course a safety issue.
Reliability wise, the 1000 cars are breaking down often and should be replaced soon. Based on limited studies, the useful life of an aluminum shell car is about 37 1/2 years. These cars did get a mid-life rehabilitation in the early 1990s, so that may extend that life some. The Rohr cars were built starting in 1976 I believe, so that time frame is probably about right (2014) for replacement.
by Retired Metro Manager on Oct 15, 2009 3:13 pm • link • report
Why build them so differently in the first place? Even if you have differences in gauge, a smaller train could still fit in a bigger tunnel, right? Why not everyone go to the same voltage and similar control systems, so at least parts of the train are consistent?
by SJE on Oct 15, 2009 3:32 pm • link • report
Putting the 1000 series cars in the middle ("belly") of a consist was done primarily to make people feel safer, but in reality the particular series of railcar you are in makes little difference in a wreck. Moving the 1000 series cars gave the appearance that WMATA was doing something to improve safety, nothing more.
Again, window dressing.
The fact is, railcars (like airliners) are not designed with crash protection in mind -- at least it's certainly not high on the priority list. Airplane crashes are extremely rare, but when they happen the fatality rate is almost always 100%. I recall reading that the seat belts and/or seat anchors in airliners typically aren't even as strong as those in a Yugo (remember the Yugo?). Presumably, the thinking is that aside from a minor slow speed collision on the tarmac, if an airliner goes down, the strongest seats and harnesses in the world aren't going to save anyone. That, and every added ounce of weight means increased fuel consumption...
Railcars are the same way. There are no 'crumple zones' like automobiles have. Metro's railcars are not structurally sound and telescope in a collision, but a railcar built like a bank vault would have its own safety issues. If there is no deformation of the frame and body at all, then there is nothing to absorb some of the energy from a collision and lessen it's severity. Passengers would be thrown forward with more force than they are in a car that partially collapses. The key word of course is partially. In the June 22 Metro wreck the fact that the lead car collapsed like it did probably saved the lives of several people in the trailing cars but needless to say it collapsed far too much resulting in the deaths of passengers in that lead car.
I'm not an engineer, and I don't pretend to have the answers, but it seems to me that any cars Metro buys in the future should be stronger and designed to partially crumple or collapse in a controlled fashion during a collision.
by Retired SMNT employee on Oct 15, 2009 8:45 pm • link • report
The far more important thing is to avoid the crashes.
That's not to say we can't do better in terms of engineering, but it does suggest that there are far better places to put our energy than into the crashworthiness.
by Alex B. on Oct 15, 2009 8:55 pm • link • report
Plus, those crash-avoidance technologies are probably cheaper than airbags anyway.
Avoiding crashes is better than surviving them.
by Michael Perkins on Oct 15, 2009 10:01 pm • link • report
It should go without saying that it is always preferable to avoid accidents rather than attempt to mitigate the consequences.
That said, I have to disagree with Alex's statement that, "Even the most crashworthy railcar isn't going to survive a crash." Obviously there are physical limits, but railcars shouldn't telescope in relatively low speed accidents like those at Shady Grove and Dupont Circle. The car (and the passengers) should be able to survive low to medium speed accidents without splitting apart and riding up onto the struck car.
Maybe there are financial or technical reasons why railcars can't be designed more like automobiles with crumple zones, I don't know. One problem of course is that there is no space that isn't occupied by either an operator or passengers on a railcar (a car typically has the front and rear of the vehicle that can be sacrificed in a wreck). What I have in mind would be a controlled (as opposed to catastrophic) partial collapse of the car -- with the cab being strengthened to protect the operator.
In any case, I agree that the primary focus should be on minimizing crashes in the first place while keeping in mind that nothing is perfect and any system can fail. Even if Metro can prevent crashes like the June 22 wreck at Fort Totten (from what my former coworkers tell me, the NTSB has identified the probable cause) there is still the possibility of derailments.
by Retired SMNT employee on Oct 16, 2009 1:13 am • link • report
Now nothing can be perfectly fail-safe, because there are always really weird sources of failure, but the DC system has a major error: loss of signal can cause a "go faster" indication (which is fundamentally wrong design). BART, with the same signal system, discovered the same subtle problem, and implemented a secondary signal system to prevent this. *In the 1980s*. Metro, despite knowing about the problem, never fixed it.
The signal system is fundamentally defective without the "BART fix" because it does not "fail safe"; and this was the cause of the deadly Red Line crash recently.
This is damning. Designing signal systems to "fail safe" is really hard, but it's the most critical and basic element of train safety these days. Any "fail unsafe" problems detected are supposed to be addressed *immediately*, not 20 years later. Management appears to have simply lost sight of safety principles entirely.
"Crashworthy" is a complete waste of time and money. Getting the signal system right -- which is much cheaper than ordering all-new railcars -- is the key issue here.
Regarding track inspection (also important), London has taken to doing weekend shutdowns of individual lines (Blue Line this week Red Line next week) to do "blitzes" of inspection and maintenance.
by Nathanael on Oct 17, 2009 7:48 pm • link • report
Oh and regarding hours of operation: the Berlin U-Bahn is almost entirely a two-track system, yet it runs all night on Friday and Saturday nights.
by anonymouse on Oct 18, 2009 1:33 am • link • report
Add a Comment