Automate Ballard Link

44 Replies to “Automate Ballard Link”

1. I agree with the concept, but to point out a few things:

   12 minute frequency is unacceptable for the branches. 9 minute frequency is in-line with both ST’s original plan of trains every 3 minutes in the tunnel, and is close to ST’s current plan of 8 minutes per line. I don’t see how grade separations are relevant to improving this frequency – it is limited by signaling systems in DSTT, not in at-grade sections.

   An important appendix is missing – where to build an OMF. With new trains, you will need a new yard somewhere along the line. The only viable place is between Smith Cove and Interbay. There are 2 options here: kick out the National Guard, or the entire Michaels/HomeGoods shopping plaza. Both will be expensive and way more expensive than any other OMF we’ve made.

   1. Expensive, yes. But, still way cheaper than a second tunnel underneath downtown Seattle.

   2. While I think it’s good to think ahead, I would suggest easing up on building further than Westlake at this point. One of the main advantages about the concept is that it is cheaper than the BLE preferred alternative on the table. Those details can be assessed once an automated line gets added to an EIS process as an alternative. Simply getting that alternative into the EIS needs to be the next step or the idea won’t go any further.

      I’ll add that the current preferred alternative was a radical change from the original WSBLE Draft EIS alternatives. It’s not too late!

      1. Certainly the focus should be on reforming BLE and axing DSTT2. But look at all the difficulties ST has created for itself by not having a long-term plan and including the necessary antecedents for the “next step” in the projects it has undertaken, such as including the means to add a connection in the U-District. It’s getting better at such “future-proofing”, but we should not shrink from declaring where ST should direct its future efforts.

      2. I think it should initially connect to Westlake and they should spend what it takes to make that connection as good as possible. It should be designed to be extended farther. But I agree — that extension does not have to have happen with the initial line. Once it does get extended it can be extended bit by bit. A station at First Hill adds a lot of functionality to the line. Even relatively small trips (e. g. South Lake Union to First Hill) become worth it (especially with the train being so frequent). Uptown to First Hill and Ballard to First Hill would definitely be much better than now (or even future buses should they head that way). Each station heading towards Mount Baker adds value.

      3. For reasons I cannot comprehend, ST seems to have an institutional resistance to thinking two or more steps ahead that is so deep it is foundational. Small steps like crossovers and branch points are minimized for cost savings in the current project no matter what the implications might be down the road.

         For Automated Ballard Link to get any traction, it has to be more affordable than today’s Ballard Link. The exciting long-term prospect of continuing to First Hill and linking with both of the other lines fulfills a promise that dates back 30 years while creating network effects and supporting the city’s long term growth. That prospect does not exist with DSTT2 and it is a huge advantage, but also kind of a distraction from the single core message that there is a better way to Link to Ballard. Automated Ballard Link plan has core advantages of smaller stations, higher frequencies, fewer transfers (with West Seattle in DSTT) and lower labor costs (solving a problem for Metro and ST now, hiring enough qualified drivers.)

         After Ballard-Westlake, in some future ST4, the new automated SkyTrain-like line, with smaller, cheaper stations, could serve the First Hill route, and that same line or a separate E-W line could serve Ballard-UW, which actually probably ought to serve U Village or Children’s Hospital as an eastern terminus rather than U District Station, but I’m getting ahead of myself there.

         Or, if we want to be less creative, we could blow our budget building a deeper tunnel under downtown that doesn’t go anywhere new — unless you count going to the (only potentially former) county jail instead of our downtown transit hub. And we could force most people to transfer including folks who don’t now (Rainier Valley to UW, UW to Sea-Tac.) That would be the current (dumb!) plan. And yes, I voted for this dumb plan in 2016 like a majority of the electorate. That was the only choice we were given!

         Automate Ballard Link!

   3. The light rail scheduling does not have to be symmetrical and the frequency does not have to be divisible by 60 minutes. Once above 15 minutes, the need for the arrival times to match the exact minute from one hour to the next is just not that important. Even Link 1 Line today has slight arrival minute variations from one hour to the next!

      Plus, frequencies higher than the base are only needed when overcrowding occurs.

      Didn’t ST recently suggest that it’s station platform capacity that’s the limitation? If that’s the case, they should prove it! And that remedy may simply be more vertical devices within the current station vaults. Until ST demonstrates that there is a platform capacity problem AND that there’s no solution, they’re throwing out red herrings.

      Rather than look at the trains per hour as some sort of fatal flaw, we should instead look to see what the maximum trains per hour can reasonably be and develop proposed frequencies from that.

      1. The unaffordable or barely affordable second downtown transit tunnel with all of its carbon emissions, construction impacts, and long transfers just to achieve inferior service at much higher cost than using the tunnel we already have, may be the plan of record, but it is madness. The financial imperative is such that there ought to be an outcry to make the most of this tunnel.

         Some cheap signal upgrades that would have been possible in 1925 are probably all that’s needed in DSTT to achieve headways that are sufficient for long term demand as described in this post. If that isn’t the case, let’s see the proof. There isn’t any. There is only hand-waving and bogus claims of things being impossible, while there are counterexamples all over the world you can find on the Internet in minutes. This line is projected to open in 2039 and the ST3 vote was in 2016, 23 years earlier. I’ll bet if you informed today’s electorate about the choices, there would be a different choice than DSTT2. There should be.

   4. 12 minute frequency is unacceptable for the branches. 9 minute frequency is in-line with both ST’s original plan of trains every 3 minutes in the tunnel, and is close to ST’s current plan of 8 minutes per line. I don’t see how grade separations are relevant to improving this frequency – it is limited by signaling systems in DSTT, not in at-grade sections.

      I agree. I think this is a flaw in an otherwise excellent essay. I have no idea why the author gave the impression that four minutes through downtown is some sort of limitation. I think he got it backwards. The only reason the trains will run every four minutes downtown is because the trains will run every eight minutes on each branch (not the other way around). The reason the trains will run every eight minutes on each branch is because they don’t want to run every six minutes down Rainier Valley.

      But three minutes through downtown is definitely possible. This has been the case for a long time as noted in this interview in 2015:

      I spoke with Marie Olson, Sound Transit’s Link Transportation Manager for Operations. As it turns out, Sound Transit’s signaling system is designed for a minimum 90 second headways. Ms. Olson believes that if the demand were there, 3 minute headways are achievable without significant deterioration in trip quality.

      Going more often then every three minutes is also possible (as she mentions):

      Although going below 3 minutes is possible, due to the variability inherent with human factors and surface operations it “wouldn’t give our ridership as reliable a service.” The small windows to fit in delayed trains might cause them to bunch up, delaying riders. Furthermore, it would likely require additional investment in Traction Power Substations.

      Thus even three minutes is not an absolute. The trains could probably run every two minutes thirty seconds without too much trouble. That would mean running every 7.5 minutes on each branch.

      Trevor Jones does make this clear (in this essay):

      Sound Transit believes that achieving better than 3-minute frequencies will require upgrades to DSTT1, including egress improvements (stairs/elevators/escalators) and better signaling. Even if these upgrades cost several times more than the 2015 estimate of $20 million, it would still be far cheaper than a new $2+ billion tunnel.

      I agree completely. We can run each branch every nine minutes easily. We can run each branch more often with a bit of work.

      [Note: As an editor I feel bad that I’m critiquing this now instead of suggesting an edit before it was published.]

      1. Link increased to 6-minute peaks in the mid 2010s temporarily between around 2012-2018, or maybe a narrower period 2014-2016, to balance ridership/vehicle constraints in the run-up to U-Link and Northgate Link. Ridership was increasing but the U-Link and Northgate Link vehicles hadn’t all arrived yet. So ST went with higher-frequency 3-4 car trains instead of lower-frequency all 4-car trains. It said this was temporary and would revert to 8-minute peaks in ST2. It reverted back to 8 minutes maybe when U-Link started or a couple years later. I think I remember taking Link at UW Station first with 6-minute peaks, later with 8-minute peaks, but I may be misremembering and 6-minute peaks was all before U-Link.

         So ST did run 6-minute peaks in Rainier Valley for some five years. Therefore it could do so again, or do it all day. SDOT long maintained 6 minutes was the limit in Rainier Valley to avoid throwing off cross-traffic cycles and pedestrian crossings. But then in 2024 it said SDOT may now be more flexible and allow 5-minute maximum. Then 5-minute all-day service got into the Ballard operating plan. So apparently SDOT is OK with it. I assume 5-minute service is to cushion the transfer blow and to avoid the glaring underuse of a tunnel with only one 8-10 minute train in it.

      2. I agree that 12 minute frequencies are unacceptable. I originally wrote 9 minutes, but Mike pointed out that the current plan is for 4 minutes combined downtown. I think ST should aim for 2:40 combined at first. 3 minutes seems like it should be possible without upgrades. My note about at-grade crossings was relating to service reliability as discussed in the interview with Marie Olson. Perhaps I should edit the post to make myself more clear.

2. This is a well-written, comprehensive distillation of the consensus which has emerged over the past two or three years here on the Blog about “BLE”. Thank you, Mr Jones. I hope Counselor Balducci reads it and puts it in front of her fellow Sound Transit Board members.

   I particularly like the path you would take with the extension south of the southern First Hill station. Serving 23rd and Yesler / Jackson is a big win for access throughout the Central District. It does complicate the interaction with Line 2 at Judkins Park, but improves the overall utility of the extension versus the original proposal just to follow Rainier.

   Also, I note that you would site the new Westlake Station south of Pine. That is good since it extends the protected access through the station southward toward Midtown’s big buildings, going some way toward replacing the lost Midtown Station.

   I do wonder, though, why you would make the station boxes three hundred feet long instead of two hundred. If the line is to be automated, you can run trains as often as does the Sea-Tac People Mover. Two-car length trains should be adequate.

   I have two addenda to bring up. First, there is some recent movement toward considering a “fishook” extension of BLE to Fremont, Wallingford and the U-District as a final step after the First Hill / Upper Rainier Valley extension is complete. So the station orientation in downtown Ballard should be rotated to east-west which will force a crossing of the Ship Canal under Salmon Bay. That would be more expensive than a tunnel under Fourteenth, but would also result in a better station location for Central Ballard. You might add a reference to the recent discussion to the list of references.

   And second, a “small maintenance facility” in Interbay is fine for aesthetic and rider comfort maintenance, but you don’t want a duplicate heavy electrical shop and wheel grinding for the Light Metro equipment. So a non-revenue connection between the new line and The Spine is necessary.

   It can be a single-track connection between Stewart and Westlake and Third and Pine. Demising a hole in the north wall of the station box at Third and Pine is much easier than connecting to a bored tube.

   The change in power delivery can be made by using a “donkey” with overhead power which can travel a but more than its own length into the connecting tunnel in the block between Pine and Stewart under Third. The tunnel beyond the curve would have third rail only but the block under Third would have both sources of power. The Light Metro cars would be automatically delivered to and retrieved from the connection tunnel.

   Given that three lines in DSTT1 requires modifications to accommodate greater rider volumes to the four downtown stations and widening the platforms is not possible, center platforms with ADA-compliant ingress / egress will be required. That means no cross-over would be possible within the Symphony Station bus bypass lane. Therefore access to the connection tunnel by the Donkey and any cars it would be retrieving from or delivering to it will have to occur during a night-time lull in service. That need not be very long — basically the time it would take for the Donkey and its train of cars to be delivered to move against traffic in the southbound tube from the first set of cross-overs south of the south tunnel portal at Fifth and Washington until it was fully in the transfer track.

   Such a delivery movement would typically be scheduled around 2 AM when headways in The Spine might be every half-hour for each line. That gives the out-of-direction move ten minutes to get to the transfer track refuge.

   The return journey with traffic can just be slotted in between in-service trains.

   1. REALITY CHECK – given the reality of what is possible north of IDS on the 1 line you will be lucky to see 5 minute headways. ST have been conducting tests trying to achieve 4 minutes and they all fail. Problems preventing the desired 4 minutes include unreliable LRV’S, Security issues, medical issues, door issues and of course operators operating at different speeds. This does not include the common issues with the signalling system and track defects. Trust me, I work in the Control Room.

      1. So fix the problems. There is no fundamental reason why the trains can’t be a lot more frequent. That is not my opinion, but the opinion of Sound Transit’s Link Transportation Manager for Operations (at the time). https://seattletransitblog.com/2015/03/21/capacity-limitations-of-link/.

         It is like saying my car can’t go over 40 because the tires are old. Get new tires!

      2. Fix the problems? Ok Ross, let’s get rid of the addicts overdosing on the trains, let’s just throw the unconscious medical emergencies off the train, let’s electro-shock anyone who holds a door, let’s magically have trains arrive at IDS ON time, let’s fix the LRV mechanical issues that have existed for 15 years, let’s stop the track from degrading and finally let’s wave our magic wand and make everything perfect!

      3. Can you explain what “failure” looks like when they try achieve 4-minute headways? Are they simply unable to pass 15 trains per hour each way through the downtown tunnel, or is the problem that the headways end up ranging from 2-6 minutes due to bunching?

         Also, do you think the issues apparently preventing less than 5-minute headways are insurmountable, or are there reasonable changes ST can make? You seem to lump in the signals and track maintenance issues with societal issues that are out of ST’s control, but ST has direct control over its operators, track, and infrastructure, so at least some of the problems preventing high-frequency operations should be fixable. Or are there deeper issues?

      4. All those problems have been happening the past few years with existing service. Semi-reliable 4-5 minute trains are better than semi-reliable 8-10 minute trains. We hope ST can fix all the delays someday, but that’s not a reason to hold up the full 2 Line.

3. The OMF does not have to be large. It’s not like this line will be 40 miles long. There isn’t a need for driver breaks either.

   If two cars ran every five minutes and the round trip is set at 35 minutes (it would actually be about 25 if it’s 11 minutes each way plus some reversal time), that’s 7 trains. With two cars per train and 3 spare trains, that’s just 20 cars to fit into an OMF. Surely, that could be sited in Interbay.

   I don’t see the OMF as a deal breaker.

   1. Agreed, though if the extensions possible were built, the number of cars necessary to serve the system fully would obviously increase.

      But I don’t think that Sunny is looking widely enough at the possibilities in Interbay. The MF can be west of the BNSF tracks; the access trackage can cross above them at the south end of the Interbay Yard throat tracks.

      There is a large, usually empty truck parking lot west of the BNSF tracks just north of the Magnolia Bridge which would be a perfect site.

      1. Every time the proposed automated line comes up, the question of the OMF follows, and the obvious site is the National Guard base.

         The National Guard is vacating the base. Possible reuses of the site have been bandied about since 2018.

         An OMF could be built on the ground, with workforce housing on top. I expect the major problem is costs to deal with the liquefiable soils covering Interbay.

      2. If the extension I proposed was built, it would be relatively easy to make at least a non-revenue connection to the existing line at Mount Baker, so the Interbay OMF wouldn’t necessarily need to grow.

      3. The soils of Interbay haven’t stopped it from being a rail yard since 1896. With all that surplus flat land the universe is trying to tell us to use it as a new base for Automated Ballard Link. That’s exactly what we should do.

         As for workforce housing on top, that may or may not be cost effective but it’s a beautiful vision. It wouldn’t necessarily have to cover the whole thing or be built simultaneously.

      4. The city has been assuming the land would go toward some use of mixed industrial/residential since 2018.

      5. Every time the proposed automated line comes up, the question of the OMF follows,

         Which is a good example of worrying about the wrong thing. As Trevor pointed out there are good options (Interbay or a non-service option) both of which would be massively cheaper than a new tunnel (from Westlake to SoDo).

4. One aspect that we don’t have the skills to answer is to determine the best place to bore or excavate a tunnel under or over the current Link tunnel. That is something that would affect station locations, transferring and even the general systems layout.

   This assumes putting it where the current BLE has it. However, as the article notes, this could result in a crazy deep tunnel. That results in crazy deep nearby stations as well.

   The prior WSBLE did not offer any alternatives other than direct crossings under DSTT adjacent to Westlake Station. ST has not analyzed other possibilities.

   For example, could the line cross above the current DSTT then drop down for the transfer station to be at Symphony?

   Or could it cross underneath lower Capitol Hill and bend back into Downtown for the transfers — or even move the transfer to Capitol Hill station? This could allow for aerial tracks above Mercer St between Seattle Center and I-5 — and that would offer a much gentler (grades) way to get shallower stations at Capitol Hill and First Hill.

   Overall, I see this underground crossing assumption as a basic question that ST needs to adequately study before and alternative concept can be promoted. ST has seemingly not yet gone into enough analysis to see if other places to cross are better.

   1. The current tunnel is not deep. There’s no way you could have another tunnel go above it.

      1. DSTT is not that deep at Westlake as you state. However Downtown is on a hill. That’s why I mentioned looking beyond the Westlake Station to see where another crossing location could work.

         A reduced centenary train clearance only has to be 12.3 feet from wire to track. (https://assets.new.siemens.com/siemens/assets/api/uuid:251fd5b2-92c9-467a-bcc1-ba6c8064a8f7/low-floor-lrv-literature.pdf)

         With only about 15 feet under an overcrossing structure needed it seems very likely that a shallow crossing could be found nearby. It’s just a matter of if it’s in a reasonable location to have a transfer station platform.

      2. Al, Light Metros don’t use catenary; they use third-rail contact and high-floor platforms. Even the freaking people mover at Sea-Tac — in service since the 1970’s — uses high-floor stations and third-rail contact.

         If we’re going to “Automate Ballard Link” the throw the damn catenary in the trash. The only reason to assume the higher costs of low-voltage catenary are that there will be people and vehicles crossing the trackway. An Automated Light Metro will certainly not be contending with vehicles crossing its path and will do its damndest to keep people out, too, except inside its vehicles. To build a new system without platform doors is malpractice.

         [Ed. Correction per author’s intention.]

   2. Al, the DSTT2 Westlake Center platforms are planned to be deep because ST wants to dive down to a fairly deep station in the International District, whichever one they choose.

      If you want to stay flat in order to go up into First Hill, the platforms can be at a higher elevation above MSL.

      Now maybe they can’t be just one story deeper than the DSTT1 platforms. There are probably supports splaying out beneath the very heavy bearing walls. But, especially if the tunnel is below Sixth, they can be shallower than ST was planning in order to lessen the grades to the south.

      1. “it” not “they” in the last sentence. “[S]tation” is singular.

      2. I believe that the planner platform at the Midtown or CID-N Station are not lower than Westlake.

         I think it’s probably more than it’s not safe to bore closer to the surface — and that the street disruption (utilities, traffic, building access) from having a shallower track would be significant.

5. Unfortunately, this means the new line would not be connected to existing maintenance facilities, so a small maintenance facility would need to be constructed in Interbay.

   Another possibility is to build a non-service connection to the existing line. This is fairly common (Toronto did it to connect the Sheppard Line to the rest of the system). This would be much, much cheaper than a full-service connection. So basically you do whatever is cheaper (and it wouldn’t be that expensive).

   1. I agree, we should do whatever is cheaper. But we might need more OMF space anyway, and it would give us more flexibility to use different kinds of trains without towing them (although it does seem like a good idea to maintain at least non-revenue capability for things like loading gauge and voltage).

      1. I disagree. One of the advantages of a Light Metro is the third-rail power, which results in a tunnel three or four feet smaller in diameter.

         I grant that third rail shoes are problematic around low-floor platforms. If we put center platforms in the DSTT1 stations then the shoes might foul the platforms. They would certainly be dangerous to passengers on such a center platform. So if ST did what I proposed for the tow moves for Heavy Maintenance, the center platforms would have to be closed during the small hours of the night. They wouldn’t be needed then anyway.

6. What I like the best about this proposal is that it can use the EIS already nearing completion for BLE, with a relatively minor addendum for a new OMF, and wherever they could figure out to turn the TBMs around south/east of Westlake. It’s easy to build less under a completed EIS, since impacts would (presumably) be simply reduced if there were less construction.

7. Station boxes shown are approximately 100 ft by 300 ft, which would only accommodate the equivalent of 3 car Link light rail trains. This seems reasonable, as the Vancouver SkyTrain uses 250 ft platforms.

   That is for the Expo and Millennium Lines. This is more than Vancouver actually needs, although they are slowly increasing capacity. To quote Wikipedia: Each five-car Mark V train will be able to hold 672 passengers regularly, both seated and standing, with a potential crush capacity of up to 1,207 passengers.

   The Canada Line has smaller platforms. They are 50 meters (164 feet). To quote Wikipedia again: The capacity of the trains is estimated at 334 people per pair of cars (comfortably) or 400 people at “crush load”*.

   To put things in perspective, Link can handle somewhere around 600 to 800 per train at most.

   The trains can carry way more people despite a much smaller platform size. Our trains are very inefficient. Each train car has a driver compartment on each end (so eight for a four-car train). In contrast the Vancouver trains don’t have any driver compartments. The trains in Vancouver also have open gangways. This increases the effective capacity as it makes it easy to move between compartments. When a train gets really crowded people are turned away at a station simply because they lined up by the wrong train car. This is less likely to happen with open gangways. Similarly the trains are high platform. This enables better flow within the train car. The capacity should be the same but if you have more doors and people aren’t crowded next to the door inside the train (because their stop is coming up) the effective capacity is higher.

   We could probably be just fine with platforms as big as the Canada Line. But we could splurge and go with bigger platforms (like the Expo/Millennium lines). There is simply no reason to have platforms any bigger than 250 feet. This makes a difference when it comes to cost as the smaller the platform the cheaper the stations.

   *Technically it isn’t “crush load”. No one is standing at the platform pushing people. But the trains are crowded enough that riders feel like they can’t get on.

8. Automated so it can be just like Kuala Lumpur, Malaysia – where the light rail just killed a blind man who fell off of the station platform. That’s where operators shine: the unexpected that no one anticipated. Otherwise, you just kill people now and again.

   1. Link light rail trains, operated by in-cab humans, have struck 52 pedestrians since 2009.

   2. Platform screen doors exist and are used extensively on other, heavier ridership SEA automated metro systems.

      1. Building an entirely new system without platform doors is malpractice. Retrofitting them to DSTT1 might be a problem since they’d take at least a couple of feet of the already narrow platforms. But to build a new Automated Light Metro without them would be foolish and cavalier toward the riders.

   3. The accident in KL was on the Ampang/Sri Petaling Line trunk section, which is the one line in KL that does have operators. The older automated line that is similar to the Vancouver SkyTrain, the Kelana Jaya Line, only has platform doors at its underground stations, but the two newer MRT lines in KL have full platform door coverage. Certainly as others say below anything new should be built with full platform door protection and the fact that we continue to build 19th-century style ‘light rail’ in the US with fully manual operations and tons of safety risks baked in is absurd.

9. Yes! What a brilliant case for what we should actually do with ST3 in Seattle. This project is currently projected for a 2039 opening. We’ve got automated taxis driving people around SF today; who thinks we can’t figure out how to run an automated train on an exclusive track by 2039? Wait, nobody, and Vancouver did this in 1986, which is before a bunch of readers of this blog (and many voters) were even born.

   The current ST3 plan for Seattle guarantees buyer’s remorse up front. Really, even if it were 100% free and could open tomorrow, the DSTT2 configuration just plain sucks, with or without breaking the Jackson hub. Hugely expensive, transfers are a pain in the ass forever, no new coverage, long, long headways. It’s not something to aspire to. And who thinks FTA is going to be funding our shenanigans anytime soon? The Mayor is spinning up a big team on this but the money isn’t really there, is it? Maybe if you cut to the bone and put the Ballard station south of the ship canal.

   There’s a ST CEO search and a KC Exec race underway. Who’s going to form an org, fund it and get the message of this article to a very broad, smart and civic-minded audience that is the Seattle electorate to stir and pot stimulate the necessary conversations? I don’t even live on this line, but I’d write a check to fund the campaign. Automate Ballard Link!

10. I hope several of you have made comments in the Ballard EIS so there a record.

11. I agree with the concept, and I believe that Sound Transit should do this with ALL of the outlying areas, such as Everett Link. I also think that the Everett Link’s next segment should be what was initially proposed, the equivalent of an above-ground version of the Northgate extension, which was to Mariner Park & Ride, which would be done a few years sooner than what they changed the plan to, which was to go to Boeing…while connecting to Swift Green to Boeing. To fund these changes, Sounder North is and has been duplicative of express bus service from all of its stations and should be discontinued, even more so since Link goes to Lynnwood now, a short bus ride from all Sounder North stations. Put the cost savings into all of the Link extensions (Ballard, West Seattle, Mariner) instead, a much better use of public money.

Comments are closed.