[UPDATE from Brian:] I went up to check out how the clean up was progressing. The LRV is now back on the rails and clear of the SB track Damage to the pantograph and skirting is pretty bad and will most likely need to be replaced. The brandt rail truck is still blocking the SB track with several men and equipment working on the rails and/or checking out the damage to the guideway. Two BNSF trucks were also around with a heavy boom/crane truck. The wind and rain is really starting to pick up. While it has taken them a long time to get this mess cleaned up and ask a lot of questions, I have to hand it to the crew that is cleaning up the incident. I for one would not want to be in it!

[UPDATE: KING 5 reports that there may be disruptions on the Tuesday morning commute. Buses may run between Stadium and Mt. Baker whenever they decide to move the train, temporarily blocking both tracks.

Martin is planning to update here at 5:30 am tomorrow.]

A Link light rail vehicle has derailed on the wye switch on the elevated section of the Operations and Maintenance facility. The cause of this is unknown at this time but it appears that the vehicle “picked the switch”. No passengers were involved and the operator is fine. This train appeared to be going out of service when I saw it. I was curious as to why it stopped! Saw it all from my window. (pic to come later)

Expect minor delays of 10 to 20 minutes to light rail service. Trains are crossing over at the O&M and using the northbound tunnel for Beacon Hill and Mt. Baker Stations. The trains will cross back over at South Walden Street.

King 5 has aerial pictures of the incident here.

111 Replies to “Breaking: Link Derailment @ O&M”

    1. I think the reaction from transit supporters should be more strident: we support good transit, but when stuff like this happens, it’s just not acceptable. This was a really depressing commute for me today.

      1. It happens. No matter what system, no matter what method of transportation, something like this happens every once in a while. Of course they’ll fix it and we’ll get back to business as usual, and maybe there’ll be a little design change in the future because of it, but it’s a lot better than traffic. :)

      2. Derailed trains that block all train service is better than freeway traffic? I disagree.

        “Link service will be temporarily suspended during the removal as the disabled train as both northbound and southbound tracks will be blocked. Passengers will be directed to back up Metro buses (Route 97), which will operate between the Stadium and Mount Baker light rail stations.”

        BTW, why complain about freeway traffic? After all, it happens.

      3. Man, I was all, what the heck could they mean. “Removal as the dissabled train?” WTF does that mean.

      4. How often do these things happen? Obviously we can’t establish a rate yet, but what’s a baseline to compare this to from other, comparable systems?

        I ride Link four or five days a week to get to work. I love it. Today, driving would really have been much better. This is the fifth day since September 9 that I wish I had driven instead. I’m generally a pretty hardcore person, so imagine how this kind of thing scares off the non-diehards. I’m willing to give it all the time it needs, but most people aren’t, and it’s probably not possible to convince them to be.

      5. I think we need to give the line more time to mature, Matt, so we develop have a sense of incident frequency. Whether this is unacceptable depends on whether it turns out to be an isolated incident, something that happens every six months, or something that happens every five or ten years.

      6. Buses break down and crash all the time, and it’s never reported. Unless the bus bursts into flames or something.

        These things happen.

      7. The difference is you can go around a broken down bus and one bus is a much smaller loss in capacity than one train. This derailment isn’t a big deal but it points out the single point of failure issue with high capacity rail. Accidents happen, engineering failures occur and in today’s world single points of failure become high value targets. Anything that can be done to make a system more robust is extremely important. That includes having back-up bus routing planned in advance of disasters be they natural or man made.

      8. Exceptionally slowly, and after exceedingly long waits. I was on one – over an hour from Westlake to Columbia City.

      9. On the other hand, you really can’t expect there to be no difference in service levels or qualities when the track is a half capacity for a segment.

      10. I expect the disrupted service to be similar to what they tell me the disrupted service will be like – 20-minute frequency – and it was not.

      11. Matt, that’s something you need to take up with Sound Transit. Let them know you were displeased with the lack of accuracy in their delay estimates. They’ve already acknowledged that they could have done a better job in getting information to the people waiting at the station platforms.
        I had just come on duty for a southbound trip, not too long after the accident, and I was practically mobbed by people on the platforms from the Beacon Hill and Mt. Baker Stations, wanting to know what was going on. Yes, mobbed, even though I was in a sealed cab!
        They were angry not only about the delay, but because I couldn’t “take two minutes to explain to them what was going on”.
        First off, I just came on duty and didn’t know what happened other than we were running single track operations… I was thrown into the thick of things without really knowing what was going on.
        I couldn’t take two minutes because it’s my job to keep the train moving along, not be public relations for Sound Transit. I immediately contacted the control center and informed them of the angry crowds and asked for them to make a station public address announcement, so people would know what was going on. Whether they did or not I have no way of knowing.
        I also informed my passengers via PA that service delays were caused by single-track operations between Mt. Baker and SODO stations, and while I didn’t know how long the delays would be, I did know that the train would be continuing as soon as clearance from Link Control was received. We couldn’t give time estimates to passengers because we didn’t know. All the operators could do was wait like everyone else, until they were told they could proceed.
        When the control center allowed communications other than emergency transmission or breakdowns, I was in constant contact with them, asking questions like, “is there a bus bridge in place between the stations?”. I wanted to inform my passengers about all their possible travel alternatives.

      12. Sounds like you were in the dark, too – the higher-ups could have done a better job communicating with operators, too, it sounds like.

    2. No it doesn’t! This derailment in the first five months of operation on a 100-year system proves transit is doomed to fail forever!

  1. Not much of a derailment, but you have to wonder how this is even possible. Trains only go about 2 mph during entry and exit from the stock yard.

    1. Thanks, Charles. The last three photos show that the wheels of the rear truck of the rear car appear properly railed.

      Most splits end in a derailment, although that’s more a function of speed than direction. This car was moving very slowly and apparently the operator noticed in the mirror that the trailing car was not tracking properly and stopped before it was pulled off the tracks. Picks usually happen at an old switch machine with some “play” in the cross bars connecting the points so that a point isn’t snugly against its running rail, when the points have been worn or broken so they’re shorter than they were when new causing a “wide spot” that slaps the wheels against the point just where it starts, or with a car that has old wheels with a thinner flange than normal. It usually takes two of the three situations working together to cause one.

      These are obviously new switch machines, new points, and new wheels on the vehicles, so the common causes of a pick are not present.

      The fact that the first five trucks on the train made it through the turnout properly shows clearly that the switch was lined for “divert” and that the points were firmly against the through running rail at least during their passage.

      When a split happens on a diverging movement, the wheels on the “through” side of the truck are able to squeeze between the through side point and the running rail. The point is supposed to be snugly against the running rail. What happens after the through side wheels squeeze behind the point is that immediately their backsides are constrained by the through side point pressing against them since the normal gap between the running rail and the point on a through movement is not present. They start pulling the diverging side wheels off the diverging running rail which is getting farther from them every foot.

      Depending how wide the gap between the diverging side point and running rail is, one of two things can occur. Normally on freight railroads the gap is narrow enough that the backsides of the diverging side wheels press against the diverging side point before the diverging side wheels are pulled off the diverging running rail. It becomes a tug of war between the wheels on each side of the truck and one or the other quickly wins, dragging the opposite wheels up and over the point constraining them. Whichever side wins it is possible but not frequent that the wheels on the side that loses will run a few feet with the flange on top of the rail and then drop down on the “proper” side, properly railed for the movement of its running track.

      However, if the gap is wider than the width of the wheel on the track, the diverging side wheels may be dragged entirely off the diverging running track and fall into the gap. That greatly increases the area of the wheel slammed against the point when it does contact the point seconds later overwhelming the now much smaller force on the backside of the flanges of the through side wheels. They are then jerked upward out of the skinny gap between the point pressing (weakly) on the running rail and will typically fall into the widening space between the running rail and the through point. From this point on, both sets of wheels are on the ground.

      Light rail wheels are typically considerably thinner than freight wheels because they carry less weight and have shorter flanges because they have to pass through gallery track in streets (or stations shared with buses in this case). I have not read that this increases the likelihood of the “unpicked” wheels falling into the gap, but I expect that it does.

      If a switch is thrown underneath a rail vehicle the trailing truck will still be properly railed for a while, until the distance between the two routes becomes greater than the distance between the trucks of the car under which the switch was thrown.

      Given these tendencies I think I’d put my money on TheSwitchWasThrown in the Fifth. Which means someone in the dispatching center likely was distracted and set up a following through movement before the out of service train had cleared the turnout.

      It also means that the turnouts themselves are not properly interlocked. Articulated LRT’s have trucks often enough that sensors on the turnouts on revenue trackage should ensure that they can’t be thrown when a vehicle is within their “plant” regardless of requests from the dispatchers.

      I hope I’m wrong, but it doesn’t look good.

      1. In the paragraph beginning “However, if the gap…”, I should have said “They are then jerked upward out of the skinny gap between the point pressing (weakly) on the running rail, ride on the flange along the top of the through running rail, and eventually fall into the widening space between the running rail and the through point. From this point on, both sets of wheels are on the ground.”

      2. I was going to post a “How did this happen” question today, but despite my general lack of reference save my own Ho, Flyer, and O Guage sets in my Grandfather’s basement, I think I understand.

        Boiled down:

        -likely cause: dispatcher error in remotely resetting switch before the out of service train cleared the switch

        -contributing factor: failed interlock/bad design that didn’t include an interlock that would prevent switch reset prior to second (or presumably third, fourth, etc.) car clearance of switch.

        While I agree it “doesn’t look good” from the standpoint of either human error or engineering; having idenfiable causes means being able to identify solutions, and as a union brother to the rail operator – I’m relieved to hear that in your assesssment operator error wasn’t likely a factor, as if anytyhing the operator was going slow enough and was aware enough of his vehicle to prevent worse damage than what occurred.

      3. I would be shocked if the turnout wasn’t properly interlocked. That’s really rather cheap to fix, however, so although it would be a terrible design error, it wouldn’t be a hard one to fix.

        I’d bet on a mechanical defect in the switch machine instead.

  2. hmmmm you have two people with the nickname “Charles”? I did not post that comment at 16:29:44.

    Guess I need a new nickname on this blog?

  3. If the derailed car was going out of service, the picture from KOMO is consistent with the switch going back into its regular position before the rail car had completely passed over the switch. The rear wheel truck is on the mainline track, while the rest of the train is headed down the side track to the yard.

    Is this what “picked the switch” means? Oh yeah:

    QUOTE Split the switch: The result of a freight car’s trucks following opposite rails (i.e., one follows the main line the other the diverging track), the car is said to have split, or “picked” the switch.UNQUOTE

    Maybe this can happen even if the switch stays in the correct position for the shunt off?

    1. That’s exactly what it looks like. I used to intentionally do this on my model railroad for kicks when I was a kid. ;-)

    2. I saw this happen on the Waterfront Streetcar one time. A car moved partially off the passing track when he needed to wait, realized his mistake, and kicked it into reverse. But the switch stays in the same position the whole time, so the car ended up straddling the space between the tracks.

  4. The derailment isn’t “that bad”

    There are a lot of ways it can happen but I won’t speculate on a public forum for obvious reasons.

    My best guess judging by how it happened is broken rail.

    NB/SB traffic is stopped right now, looks like there is some equipment up there now. How long it takes to put it back on the rail, inspect the area and open the gates is unknown.

    1. It doesn’t seem bad compared to other derailments, but compared to not derailing, it seems pretty bad. Plus a heck of a lot of us were stranded without hardly any information. I understand that the kinks are being worked out and all of that, but this just looks like amateur hour from the perspective of a daily rider.

      1. The information thing is an issue that will get a lot better at the end of the month. :)

      2. What happens at the end of the month? Will the operators be trained to speak directly into the PA system microphones at that point?

      3. Haha, I hope. But the end of the month is when they are planning to have real-time arrival information displayed on the in-station signs.

      4. Of course that’s not quite sufficient in an event like this, where there are exceptionally long delays and passengers need to board on the other platform at some stations. I can understand real-time arrival information taking time to implement, but can’t they manually change those signs to say something more useful than “Welcome to SODO” in this kind of situation?

        On the other hand, I’m subscribed to email alerts from Sound Transit and got an email at 3:46 saying “Trains run every 20 minutes due to train interference. Board Link on northbound platform at Mt Baker, Beacon Hill, SODO.” So at least tech-savvy transit nerds like me knew what was going on.

      5. I’m a tech-savvy transit nerd, too. But just because I’m savvy with technology doesn’t mean I have it with me on the train!

      6. Unfortunately, the “mobile version” of the update (which I had sent to my non-browser dumbphone) cut off the part of the message that told how long delays would be and to bar at NB platforms.

      7. “Amateur” would not be the most accurate word. Inexperienced, yes. It’s a new system, and it’s a new experience for just about everyone from operations staff to passengers. I will not speculate about the cause of this particular failure, because I don’t know, I wasn’t there.
        I will say that it’s my personal opinion that the ‘weak link’ is not with the operator or the training we receive. Our training is very comprehensive, and operators are proficient in their job, otherwise they wouldn’t still be there.
        The fact that all light rail operators were chosen from the ranks of bus drivers means that their situational awareness skills are more finely tuned than those of the average person on the street. This gives me much greater peace of mind when I’m off duty and riding light rail as a passenger.

    1. Ugh. So many people don’t know what they’re talking about, someone was under the impression that this was the first run ever.

  5. According to the rider alert on the ST website:

    Updated Date:11 / 16 / 09 – 4:11 p.m.

    Due to a disabled train in the SODO area, Central Link light rail will operate every 20 minutes until the end of service on Monday, November 16. Please board Central Link on the northbound platform at the following stations:

    * Mount Baker Station
    * Beacon Hill Station
    * SODO Station

    1. I love how they avoid the word “derail.” The train is disabled because its wheels aren’t touching the tracks!

      My Link operator kept referring to it as a “blockage,” but then again she also referred to the main dispatcher as “the man upstairs.”

      1. Like I said, some of its wheels weren’t touching the rails! I saw it firsthand from a passing train.

  6. Why are they not running shuttle bus service? I know they do this on MAX in Portland when there is an issue.

      1. It would run from Stadium station since the easiest way to get to beacon Hill station is via beacon Ave and Mount Baker or Columbia city station is where the shuttle would terminate (I don’t know where the cross over is.)

      2. Trimet runs buses as soon as any train breaks down, regardless of the incident. I get news alerts regarding it all the time.

    1. They will be running shuttle bus service when they have to close both directions of track. For right now trains are still operating, just on reduced frequency.

      Here’s an update to Sound Transit’s rider alert:

      “Link service will be temporarily suspended during the removal as the disabled train as both northbound and southbound tracks will be blocked. Passengers will be directed to back up Metro buses (Route 97), which will operate between the Stadium and Mount Baker light rail stations.”

    2. This was on the ST site, too:

      “Link service will be temporarily suspended during the removal as the disabled train as both northbound and southbound tracks will be blocked. Passengers will be directed to back up Metro buses (Route 97), which will operate between the Stadium and Mount Baker light rail stations.”

      My wife read this to me when I was stuck at Stadium, and it sounded to me like that was happening right then. Now getting home and reading everything, I understand that Route 97 will be happening only while they’re removing the train, but at the time I was all, “Where’s my Route 97?!”

  7. Judging from those Tweets, it sounds like the 20 minute delay was WILDLY optimistic. Sucky night to be stuck at one of the outdoor stops waiting for a train…

    In the nearly 4 months since the line opened, I can think of at least 3 major disruptions to service:
    — The tunnel closed for over a day with signaling issues
    — Motorcycle crash on MLK when SPD kept the tracks closed well into the AM commute
    — Derailment

    Stuff like this is bound to heppen, but it doesn’t seem like ST is getting any better at reacting when things go wrong.

      1. One thing you cannot avoid compleatly is Murphy’s Law. thingswill always go wrong. Sometimes it happens even though you do everything right.

    1. I waited 23 minutes for a southbound train out of Westlake Station. I arrived at the station at 5:24 and the train didn’t arrive until 5:47. I got off the train at Mt. Baker station at 6:10. I don’t mind the delay – these things happen – but I was disappointed that there was no supervisor or notice in the station to let riders know what was going on. The train was packed with confused commuters.

      1. At the end of the month, the readerboards will be operational, and you’ll get more information that way.

      2. Sure… But in the meantime, a supervisor or two in each station to let folks know what’s going on would be appropriate.

      3. I generally agree with your response to the incident, Ben… but this particular point is simply unacceptable.

        When the train became disabled, ST should have immediately called on all of its conductors and transit police to travel to the nearest station to help direct riders and answer questions.

        That’s all that would have been required to make the situation a little better.

        It would be interesting to learn if such protocols were followed and how many staff were available at that particular time to help out…

      4. The sign boards are already functional, just not with next train information. I was in Pioneer Square Station during the incident and there was a generic “experiencing delays” message. Then on the train we stopped at Stadium and the driver announced there was a derailment and there would be an “unknown” delay.

        The generic message isn’t really useful. It was well known to people online that headways were dropping to 20 minutes; some sense of the order of magnitude of delay tells me if it’s worthwhile to go above ground and find a bus or not.

    2. Sound Transit isn’t exactly good at telling people they screwed up or are in the process of unscrewing themselves. They’re really good at sitting around and talking bout stuff. Probably has to do with all the crappy politics they had to endure years ago.

  8. Brian,

    To have ended up like it did, the rear truck of the rear car had to have picked the switch. Or, and this would be a serious interlock failure, did the switch throw to the through routing under the train?

    1. From the pictures and what I can see I would venture a guess that it did throw the switch underneath after one set of wheels went over the switch. Again, not fact, just my opinion by looking at the pictures. If thats the case then they need to take a serious look at their interlocking machines.

      1. Some people here have suggested that the trains’ slow speed entering this ramp would have precluded a derailment, but a slow-moving train would be more likely to have a switch flipped underneath it between trucks, no?

  9. Sound Transit NEEDS TO LET PEOPLE KNOW WHAT’S GOING ON I was at one station and nothing on the screens saying DELAYED!!!!!!!!!!!

    1. I agree, when there are going to be delays the riders have a right to know. If delays are too long than changes to travel plans need to be made.

  10. Given the choice, I’d much rather be in a stranded train and have to transfer to a shuttle bus than be sitting in my disabled car on the shoulder of a freeway, waiting for a tow truck.

  11. “Meh. It happens.”? That is pretty cavalier for an elevated train. Had there been people on the train people could have been hurt and had the train left the elevated platform people could have been killed.

    Several years ago the monorail caught fire and there was a picture of a woman hanging out the door holding her baby. That woman is a friend of mine, she was travelling with two children and had to entrust one of them to a stranger. I can tell you that she would not be saying, “Meh” about the train derailment. My friend has always been a proponent of mass transit and ST as just lost one potential rider.

    1. You’ll want to stay off the “free”ways too – mayhem out there, especially in stormy weather!

    2. There are guard rails to help keep trains on the elevated guideway, even in the case of derailment.

    3. The big difference between Link and the monorail is that in an emergency people can evacuate the train via the elevated structure. That option is not available on the monorail, people have to wait to be rescued by another train or the fire department.

      Your friend would be dumb to not commute by train as it is the safest mode of transportation. Does your friend risk her childrens’ lives by driving with them? Driving is one of the most dangerous activities that people do, yet people choose to do it every day. In the US, motor vehicle collisions are the leading cause of death of children 9-18. By 2020 the number of deaths worldwide from auto accidents will exceed deaths due to HIV/AIDS. So much for common sense.

      1. Light rail requires egress from elevated structures precisely because this sort of incident is expected. Yes driving is more dangerous than riding on a train. But most commutes, at least from the suburbs are going to involve an auto component anyway. Given the “Meh, it happens” and sure we’ll sound proof your house response (didn’t want to sit outside in the summer anyway did you?) it’s hard to believe any of the promises ST presented to voters in order to get this passed.

    4. When I lived in Chicago, an elevated train derailed and fell to the street below. People were injured – some died. The train was going too fast around a curve.

      I’ve been riding Link between Beacon Hill and Pioneer Square almost everyday since it opened. About two weeks ago the train came out of the Beacon Hill tunnel and came to a quick stop on the elevated portion where the rail curves to align with 5th Ave. It felt like it was going to derail.

      There seems to be a design problem aggravated by operational issues. “It happens” just doesn’t cut it and I’m not willing to risk it while ST works out the bugs (or doesn’t). All the people who think that “It happens” is an appropriate response to this should ride light rail everyday for the next few years.

      1. The Chicago els were in ungodly awful shape for a very long time — and are very different construction from the Seattle elevated. Actually flying off an elevated line is very different from a derailment on an El with escape walkways and guardrails.

        I certainly think they need to find the source of the problem, but I am pretty sure they are going to — protocol requires a full investigation for *every* incident, unlike with cars.

  12. Last rider alert update on Sound Transit’s website was way back at 4:11 p.m.

    I am very disappointed with their poor customer service.

  13. Hey light rail supporters, you can still criticize a transit agency or light rail line while still being a light rail supporter. Just a heads up.

  14. I heard that no one was on board or at no one was hurt. If this is true than thank God for small favors.

  15. And as we can see from the photos, not just a little! The driver thought everything was going so well- until the rear of his train started to catch up with him on the parallel track.

  16. I flew back into Sea-Tac last night and took the light rail from Tukwila to the International District, boarding at 9:45 pm. Let me share my own experience. It was not a “minor delay of 10-20 minutes.” The trip to the International District took 1 hour and 20 minutes rather than the usual 30 minutes. The train would just stop for extended periods without moving in the Rainier Valley. There were no announcements of what was going on. I only found out what was going on by calling my wife on my cell phone and asking her to check the news. I pressed the intercom button on the train to call the operator and ask what was going on, and no one responded. There was no alternate bus made available, we were just stuck on the train. While I understand this was an accident, Sound Transit really needs to do a much, much better job of communicating with its ridership and reacting to a contingency next time if it wants to attract and retain riders.

    1. I agree that Sound Transit did a poor job of getting information to passengers. You obviously weren’t riding my train, or you would have heard PA announcements in addition to the ‘canned’ variety.
      Everyone seems to be under the impression that the operators know everything that’s going on all the time… that’s not the case. At one point we were instructed to not use the radio unless we had an emergency or the train breaks down.
      My PA announcements were that service delays were being caused by single track operation between Mt. Baker and SODO stations, that is both northbound and southbound trains taking turns using the northbound track, usually two trains in one direction at a time. I couldn’t give time estimates of the delays involved because we had to wait like everyone else until the control center gave us clearance to proceed.
      I didn’t withhold that a train had derailed causing a blockage, I didn’t know… at the time all I knew was that there was a blockage on the southbound track at OMF and that we were doing a single track operation in the affected area.

      1. Are you certain your PA announcements can be heard? My operator clearly wasn’t – we’d hear the chime, then some mubles and scratching sounds. They should have an automated message before an operator announcement that asks for complete silence, and the HVAC system should shut off so we can hear. Or the volume could be higher!

      2. New system, new problems. Bugs to work out. In the case where Gary called his wife to have her check the news, he probably ended up knowing more about the derailed train than the operator did at that point. The control center does not always explain why we’re doing something, they just give us instructions and we follow them.
        I was nearly mobbed by angry passengers at Beacon Hill and Mt. Baker Stations wanting to know what was going on. I had no information to give them because I had just come on shift. I was thrown into the thick of things without knowing what was going on. I requested that the control center make a station PA announcement if possible to let people know what was going on.
        I know my PA announcements are heard because when I get on the mic at Tukwila to tell someone that “the doors are enabled for their use, push the red button on the door to open it”, they go right for it… but I think they believe it’s going to bite them or something, because the first push or two is usually too light, and the doors don’t respond! : ) Look for the green LED lights lit up around the red button… that means the doors are enabled. We don’t leave the doors wide open at the terminals during bad weather because it’s cold outside.
        A tip on “canned” announcements: When the announcement is that a train is being held “due to traffic ahead”, if it’s in the Downtown Transit Tunnel it means there is a bus at the platform ahead, and the train is waiting for the bus to leave. Out on MLK it’s usually to keep interval spacing between trains. SDOT has complained to Sound Transit that trains mess up the traffic signals when they bunch up on MLK… so the control center tries to keep trains at least 7 minutes apart, especially during rush hour.
        “We are experiencing a service delay” is a more general announcement, just meaning that the trip is going to take a little longer than anticipated.
        Please note that the intercom to the operator is for emergencies only. If you have an emergency and don’t get a response right away, re-push the button and speak to the operator, stating the nature of your emergency.
        Operators are becoming apathetic to the emergency intercom calls because 99.9% of the time people are pushing that red button on the bottom of the spring-loaded seat thinking it’s the seat release. Operators generally just cancel the calls, because they’ve answered so many calls that have turned out to be mistakes.

Comments are closed.