I-90 Interchange Improvement Project & West Station | Allston

Re: I-90 Interchange Improvement Project (Allston)

I'm starting to get concerned, for instance, about the number of at-grade intersections a grid would impose on the Pike access points from Cambridge St. some of that can be addressed with signal arrangements and permitted turns, but four or five stoplights to reach the right turn into Cambridge is not an acceptable constraint to place on that movement, particularly as transit access into Cambridge from the West is really crummy until the Urban Ring comes online in 2175. MassDOT has hidden a lot of those intersections from their analysis because they aren't going to build them - Boston will.

Meh, the Pikes ramps are buried in the grid in the Back Bay, and it works allright. The arrangement in Beacon Park, even after full build, is far more logical than that mess too. Of course, the DUH obvious solution is a street at Malvern/Alcorn/Babbock, pointed out by just about everyone in the community, but MassDOT doesn't want to bother BU and keeps using a bullshit ADA excuse for why they can't do it, which I've already proved false.

persp-malvern-st.png

^That's 23'6" of clearance from ground level to the bottom of the bridge, at a 5% grade, folks. More clearance and a gentler grade than required.
 
Re: I-90 Interchange Improvement Project (Allston)

Meh, the Pikes ramps are buried in the grid in the Back Bay, and it works allright. The arrangement in Beacon Park, even after full build, is far more logical than that mess too.

The Back Bay exit is primarily designed to serve the Back Bay and South End, a catchment that's far smaller than Beacon Park's and one that is centered on the ramps. It also has ramps that merge directly into major arterials as opposed to hugging the highway a half-mile from them.
 
Re: I-90 Interchange Improvement Project (Allston)

The collector-distributor roads that DOT has designed will handle the traffic just fine. I actually find it amusing that they're seemingly importing the design from Houston, where it's practically standard procedure. And yes, they're modelling the intersections, for whatever that's worth (I think it's worthless, but I think all 20+ year models are worthless). Actually, what they're not modeling are the roads that are supposed to be built by Harvard -- Stadium Way and East Drive -- that will make everything work really easily like a grid is supposed to do.

Really, the main thing I'm worried about from the modeling is that they will force us to accept unnecessarily large roads that will be pedestrian unfriendly and make development of the parcels much more difficult than necessary.
 
Re: I-90 Interchange Improvement Project (Allston)

I made a quick 3D schematic to show how access around the new ramps would work. It's not really to scale and obviously the Pike and ramps would be curved, but it's easier to see how circulation around the project would work this way.

15209235958_09f2ef33fa_o.png
 
Re: I-90 Interchange Improvement Project (Allston)

The collector-distributor roads that DOT has designed will handle the traffic just fine. I actually find it amusing that they're seemingly importing the design from Houston, where it's practically standard procedure. And yes, they're modelling the intersections, for whatever that's worth (I think it's worthless, but I think all 20+ year models are worthless). Actually, what they're not modeling are the roads that are supposed to be built by Harvard -- Stadium Way and East Drive -- that will make everything work really easily like a grid is supposed to do.

Really, the main thing I'm worried about from the modeling is that they will force us to accept unnecessarily large roads that will be pedestrian unfriendly and make development of the parcels much more difficult than necessary.

I'm not concerned with the C/D roads, I'm concerned with the interstitial grid between the highway and Cambridge St. As far as I know, MassDOT did not model more than the 1 parallel road between the two, and if we're not going to have superblocks there would have to be at least 1 or 2 more roads than that intersecting each of the access roads (Seattle and East in Dave's drawing).

I realize that MassDOT assumes that the roads from the EB C/D road descend too slowly to have any intersections other than that one parallel road, but both of the BSA concepts assume that they do (by changing the elevation level of the site to match the roads). At some point, you're making a tight grid. When you get off at BBY, that's not really an issue because you probably aren't going any farther than that neighborhood. When you're using this exit to access places like Porter and Longwood, asking people to creep through a dense urban fabric to get out to Cambridge starts to be a little problematic.

I'm not advocating suburban-style design, and I'd like to see a dense urban fabric here for lots of reasons, but we also need to stay cognaizant of how important this project is beyond the immediate neighborhood from a connectivity standpoint.
 
Re: I-90 Interchange Improvement Project (Allston)

The drawings that I've seen typically have 2-3 north/south connectors from the C/D roads to Cambridge Street (and beyond). The DOT folks have even tossed in an extra link to SFR in order to alleviate that concern.
 
Re: I-90 Interchange Improvement Project (Allston)

Today's press conference "announces" West Station

Quick facts gleaned from Twitter:
-Harvard will help pay for it
-4 tracks, 2 platforms: one to North Station, and one to South Station
-Will have DMUs

Maybe it's time for a "West Station" topic? I'll be adding to this as more info comes out.
 
Re: I-90 Interchange Improvement Project (Allston)

Today's press conference "announces" West Station

Quick facts gleaned from Twitter:
-Harvard will help pay for it
-4 tracks, 2 platforms: one to North Station, and one to South Station
-Will have DMUs

Maybe it's time for a "West Station" topic? I'll be adding to this as more info comes out.

Link?
 
Re: I-90 Interchange Improvement Project (Allston)

From UHub (Adam was at the press conference):

UniversalHub said:
West Station will be built near the present turnpike toll plaza and should open when the turnpike there is straightened, in 2020, state Transportation Secretary Richard Davey said.

The station will be servced by DMUs, single-car diesel units that could bring subway-like schedules.

Davey said he hopes the new trains could eventually go to North Station via the Grand Junction bridge over the Charles, but acknowledged that will take considerable work, both to fix the tracks and convince skeptical Cambridge officials.

Gov. Patrick said if the Grand Junction could be put into service, that could one day lead to service from Allston to Assembly Row in Somerville.
Davey said Harvard, which owns the freight yard, agreed to kick in a third of the station's cost. The state will pay another third, with the source of the remainder to be determined.

Patrick and Mayor Walsh said Harvard has also agreed to make the streets in the new neighborhood that will spring up over the next two decades "multi-modal" - to safely carry pedestrians bicyclists and motor vehicles.
 
Re: I-90 Interchange Improvement Project (Allston)

Will this stop at Yawkey too? The articles I've read are saying it's on the Framingham/Worcester line but haven't mentioned any stops. That would be a huge win for the Fenway neighborhood... quick access to both Allston and Back Bay/South Station without having to wait an hour or two for the commuter rail (or suffer the green line).
 
Re: I-90 Interchange Improvement Project (Allston)

"Gov. Patrick said if the Grand Junction could be put into service, that could one day lead to service from Allston to Assembly Row in Somerville."

LRV Urban Ring would be a better and more likely scenario IMO
 
Re: I-90 Interchange Improvement Project (Allston)

How about somebody, somewhere show some traffic modeling that the Grand Junction can actually be put into service more than a dozen times a day before we make more kwaaaaaaaaaaaazy pwomises! Because that Worcester-NS study that Cambridge had a collective heart attack over hit its limit in gate closings at 5 rush hour trips each peak. So multiply the TPH by 3 and extend the operating hours to all-day because...magic?

Life as a lame-duck governor must be a non-stop gas.
 
Re: I-90 Interchange Improvement Project (Allston)

What are the peak DMU frequencies we could see on the SS branch? Could it approach anything close to the 7-10 minute headways we see on other lines?
 
Re: I-90 Interchange Improvement Project (Allston)

What are the peak DMU frequencies we could see on the SS branch? Could it approach anything close to the 7-10 minute headways we see on other lines?

See the discussion on p.14 of this thread about that. Short answer: yes, if we're talking DMU's in a vacuum. 15 minute headways, not 7-10 min. headways...nowhere is < 10 mins. possible out of South Station when it's at peak load feeding that much traffic into Back Bay. Nowhere did anyone promise promise subway-level headways; that's not what Indigo is, or ever could be.

It gets much more doubtful you can pull this off on the existing infrastructure when you factor in DMU's + a Framingham/Worcester schedule that's going to grow much heavier by its lonesome in the next few years.


Short answer as to why. . .

-- The DMU's are fine by themselves; there's 3 crossovers between Back Bay and New Balance that allow Worcester trains to pass. The DMU's won't be the ones delayed.

-- The problem is Framingham/Worcester and the track situation outside the scope of this DMU project. Between the west (New Balance) end of Beacon Park and Framingham station there is only 1 set of crossovers at Wellesley Hills for staging train meets. And trains have to hop over each other all day long between Beacon Park and Wellesley when the Newton stops are being served.

-- Certain sequences of Framingham/Worcester trains have very low margin for error on their schedule. It all has to do with trains serving the Newton stops vs. trains skipping the Newton stops, and the single-track platforms in Newton forcing trains to switch tracks. Every X many trains there's an extra-vulnerable slot where the track-switching dance between Beacon Park and Wellesley has to be made spot-on or things go to hell.

-- A vulnerable slot that blows its meet is going to have to pause and wait a very long time for the conflicting train to clear. That is why when Worcester is late, it is late by a mile and you routinely hear horror stories of 30 minute late, 60 minute late trains that sit at a dead stop for 20+ minutes at a switch. When one Worcester/Framingham train is late by a mile it has the added effect of backing up other Worcester/Framingham trains in vulnerable slots. There's no in-between...the line is either on-time or it's a full-on disaster commute. And this vulnerability is at its absolute most precarious at the AM and PM peaks.

-- The only way to fix this is to install new crossovers. One in Newton before Newtonville. One at Riverside Jct. A re-spacing of the one by Wellesley Farms out further into Wellesley. And maybe one more before Framingham. That solves the problem of having too few passing opportunities, lets locals and Worcester expresses coexist in very large numbers, and gives trains that blow their slots lots of recovery time. It's how every other high-traffic line on the system works, and how the outer half of the Worcester Line works.

-- The signal system is too old and limited to simply retrofit. Extremely expensive and unreliable to perform surgery on. It was considered several times in the past, and every time it projected too difficult. Basically, all 21 miles of signals from Back Bay to Framingham have to be junked and re-laid from scratch to really make it work. Requiring the same total teardown/rebuild job they're doing on the Fitchburg Line today with those weekend closures. Easily $50-75M, with zero funding for that alloted or hinted to in the Transportation Bill.

-- However, mix the full 15 minute headway Indigo service into the mix with the full Framingham/Worcester schedule (which is going to get bigger still in the next 3-5 years) and things get dicey if you don't spend the $$$ for the signal do-over. Something as inocuous as needing to hold open the doors for a couple minutes longer at Back Bay or Yawkey for a ballpark crowd or extra passenger assistance lowers the margins enough that the meets could get a couple minutes out-of-sync between New Balance and Wellesley Farms. And then...delay catastrophe.

-- Not even doubling up the Newton platforms helps much here beyond the fringes, because express trains still have to hop over locals. At most that'll shave off a few trains that won't have to be crunched into a do-or-die schedule meet. There will still be trains crunched into the timing dance between New Balance and Wellesley Farms, and the consequences for blowing those meets will be disaster commute all the same. This doesn't go away until the signals are redone.

-- This is a necessity even if the DMU's never happen. Within 5 years there'll be enough Framingham/Worcester trains that they'll start scraping up against the ceiling. A couple of Amtrak Inlands will probably force the tipping point for all the signal work $$$ because those won't be able to handle the meets either. The Worcester Line has huuuuuuuuuuuuge untapped capacity if they fixed this. Track 61 DMU's to West, DMU's to Riverside, Framingham locals, Worcester expresses and non-stop superexpresses, Worcester rush hour trains routed down the Grand Junction (the 5 per AM / 5 per PM peak plan in the prior study, not the DMU), Amtrak Inlands, Amtrak Inlands down the Grand Junction, a daily Boston-Montreal round trip. Yeah...it can handle all that if you give it nimble passing opportunities inside Framingham. It wouldn't be too too far removed from the kind of traffic levels the NEC Shoreline used to be able to handle in the slow-speed pre-electrification days. But it's gonna cost them to renew that much track infrastructure. And they haven't said a peep about how/when/for how much.



-- NOTE: New Balance and West, being inside of the Beacon Park crossovers where there's plenty of passing space, should also be fine to add to certain Worcester/Framingham schedules when they're built. Without the signal work. They have minimal effect on the Newton dance at conventional commuter rail schedules. Not enough to degrade the on-time performance to worse than the (crappy) it is now. New Balance definitely is going to pre-date the DMU's by several years anyway. West easily could too, and still be a quite very useful station at conventional commuter schedules (BU will be fetching a lot more of the patronage for the first decade anyway since Harvard land is going to take awhile to build). So keep in mind it's the extra layer of overlapping services on the schedule that's the destabilizer...not the presence of these extra station stops themselves. The track has to be reconfigured to juggle the services.



Everything that has to be done here is outside the project area of the DMU service terminating in Allston (Riverside...that DMU service they definitely can't swing without the signals). So if they're not funding the signals/crossovers now or talking about how they're going to make up that $50M+, it comes down to this:

-- You can run the DMU's. The DMU's will never be made late by a Framingham/Worcester train.

-- You can build the stations. The stations can serve the existing schedules years before the DMU's even debut.

-- ...but, are the political consequences of further damage to Framingham/Worcester on-time performance too severe to proceed with the DMU's here before they can fund the signals. The way the Worcester Line is growing, and with how critical a function it's starting to serve on horrendous MetroWest highway commutes...can it realistically happen without addressing the rest of the line? My guess is no given how much representation the Worcester Line gets in the Legislature (for good reason).

-- If that's case, when are we going to get a schedule and funding commitment for the track work? The longer that goes unmentioned, the longer you can push back the rollout of DMU's from South Station to West. This Indigo leg will start to lag well behind Fairmount and Track 61. This is an unavoidable part of the plan...answer whether anything is doable without first fixing Newton-Framingham, then quantify how much that slows the rollout to full 15 minute headways. Because this isn't going to be a well-patronized service if the headways meander at 20-25 minutes instead.




So...we're only getting part of the picture. They can't just stick a podium in front of the site of future West Station and talk about DMU's from West-inbound in a vacuum when all of the most tangible project risks and impacts are outside the city.




And no, I don't think Grand Junction service is ever going to happen. The traffic queue math around the grade crossings is fully available in the last study that got soundly rejected by the locals, and 2 minutes of feeding those timings into a calculator multiplied by simple TPH counts paints a stark picture of what a carpocalypse that's going to be at Indigo-level headways. Try it yourself...it's easy enough for even the most casual observer to do the arithmetic and see how infeasible that's going to be. I'd wager on that extra West platform getting cut from the design quickly and deferred to a later add-on while they back away slowly from that Grand Junction promise.
 
Last edited:
Re: I-90 Interchange Improvement Project (Allston)

And no, I don't think Grand Junction service is ever going to happen. The traffic queue math around the grade crossings is fully available in the last study that got soundly rejected by the locals, and 2 minutes of feeding those timings into a calculator multiplied by simple TPH counts paints a stark picture of what a carpocalypse that's going to be at Indigo-level headways. Try it yourself...it's easy enough for even the most casual observer to do the arithmetic and see how infeasible that's going to be. I'd wager on that extra West platform getting cut from the design quickly and deferred to a later add-on while they back away slowly from that Grand Junction promise.

We have been over this before, and it still doesn't pass the smell test. First off, there is no Grand Junction DMU study. There is a Grand Junction Transportation Study that did not mention DMUs because DMUs outside of Fairmount was never proposed by MassDOT until the map that went out with the CIP last year. I'm sure you did something more complex than simply "multiplying by TPH", because that study assumed commuter rail train lengths and speeds, both of which would be improved upon by DMUs.

The closest thing we have to a Grand Junction DMU study is this thesis paper from MIT:

http://dspace.mit.edu/handle/1721.1/73788

In it, the author quotes a "gates down" time of between 3-5 minutes during the peak hour for all rail alternatives, with the DMU performing (as one would expect) better than commuter rail. Now, he assumes higher frequencies than are realistic, so with 4TPH, the math is simple: 20s lead time on the gates + 15s time to clear the intersection + 10s to raise the gates, all multiplied by 4 = 3 minutes, and that's being generous with the raising time (I guessed).

You don't have to do complex queuing theory to evaluate that. It's the equivalent of a signalized intersection weighted 20-to-one in favor of Mass Ave. It's not unreasonable at all.

In fact, it's done elsewhere. Oakton St. in Skokie, IL, for example, has crossing gates for HRT: https://www.google.com/maps/@42.0255652,-87.7472286,687m/data=!3m1!1e3

The 2009 PM Peak traffic count for that intersection is slightly less than the 2035 no-build identified for Mass Ave. in the GJTS (very slightly, and the frequency is higher):

http://www.skokie69.k12.il.us/Staff/Admin/district/docs/DowntownSkokie&OaktonStreetCorridor.pdf

There is no carmageddon at that intersection. Not even close. In fact, the CTA recently added a station right after it that forces the trains to slow down as they approach, and they faced basically no local opposition in doing so (the town paid for the station).

I am not arguing that Grand Junction DMUs should be a high priority project. I am not arguing that they are cost-effective or will justify their costs with ridership. However, arguing that the equivalent of a hugely favorable signalized intersection will destroy the flow of Mass Ave just doesn't hold up. Just because Cambridge freaked out at a couple of extra minutes of delay doesn't mean they were right to.
 

Back
Top