29 December 2012

Grade Separation: The Decadal View

For the last few years of debate around the issue of high-speed rail, grade separating the peninsula rail corridor was often cast as an all-or-nothing proposition.  This extremist view clouded two important facts: first, the corridor is already mostly grade-separated (in 2013, only 40 road crossings out of 104 between San Francisco and San Jose will remain at grade); and second, grade separation is a slow and inexorable process that takes place over many decades.

If we assume the next several decades will be like the last several decades, we can take an educated guess about how and in what order grade separations will be built.  The criteria for prioritizing each project could be:
  • creating long, uninterrupted stretches of grade separated right of way to enable higher train speeds without compromising safety
  • creating a four-track mid-line overtake facility to increase the capacity of the corridor, to support initial HSR service
  • separating crossings that rank highest in the CPUC's Section 190 Grade Separation Priority List
  • delaying the most expensive and politically costly projects until last
Phase I is simply the completion of the San Bruno grade separation in 2013.  The San Mateo / San Bruno grade crossing being replaced was once rated #7 statewide on the CPUC's priority list.

Phase I: San Bruno Grade Separation

Phase II consists of four projects in San Mateo County, opening up two long stretches free of crossings by the early 2020s, including 14.8 miles free of crossings north of Burlingame, and 6.5 miles south of San Mateo.  This enables the future construction of the "short" mid-line overtake envisioned in Caltrain's corridor capacity analysis, and leaves only a few dense clusters of crossings within San Mateo County.  The new grade separations, in order of priority, are:
  1. 25th Ave in San Mateo, the only grade crossing that remains between San Mateo and Redwood City.  Along with new grade separations already planned at 28th and 31st, this project enables the future 4-track mid-line overtake.
  2. Broadway in Burlingame, an extremely congested crossing that has been slated for grade separation since the 1970s.  It rates #11 statewide on the latest CPUC priority list.
  3. Linden Ave in South San Francisco, originally planned as part of the San Bruno project, but dropped from the final design in 2007.  Grade-separation at Linden is accompanied by the closure of Scott St, which becomes a pedestrian tunnel.
  4. Center Street in Millbrae, a grade separation that will require a U-shaped elevated ramp due to the nearby BART subway tunnel box.  Such are the consequences of bad planning.
Each of these projects is independent and can be negotiated on a case-by-case basis with the four affected cities.  Starting in San Mateo County allows at least another decade for the Pacheco vs. Altamont debate to run its due course, legally and politically; these four projects are useful either way.

Phase II: San Mateo County Grade Separations
Phase III occurs mostly in Santa Clara County, creating a new stretch free of grade crossings from the southern half of Palo Alto all the way to San Jose in the late 2020s, assuming the routing of HSR over Pacheco survives as currently planned.  The last two grade crossings in San Francisco are also eliminated as part of the downtown extension project.  This phase includes the following six discrete projects:
  1. Mary Ave in Sunnyvale, the corridor's busiest grade crossing in this county, with more than 25,000 daily vehicles
  2. Sunnyvale Ave
  3. Rengstorff in Mountain View, with about 18,000 daily vehicles
  4. Castro in Mountain View, with about 9,000 daily vehicles
  5. Charleston and East Meadow in Palo Alto, with a combined ~30,000 daily vehicles plus numerous pedestrians and bicyclists
  6. 16th and Common in San Francisco, as part of the DTX project
Together, these six projects create a new 16-mile stretch of track that is entirely free of grade crossings.  The corridor is now left with just three dense clusters of grade crossings, in San Mateo / Burlingame, Redwood City, and PAMPA (Palo Alto - Menlo Park - Atherton), highlighted in orange in the figure below.  Note that these three dense clusters contain 27 crossings, and that to get this far, only 12 existing crossings were newly separated.

Phase III: Santa Clara County Grade Separations
Phase IV is the Great Redwood City Grade Separation.  This project, potentially for the early 2030s, would prolong the four-track mid-line overtake by three miles, by removing six grade crossings in downtown Redwood City.  Removing this cluster first makes sense from the standpoint of increased corridor capacity, the lowest number of new structures, the short mileage, and the entire project being politically simplified by virtue of its containment within Redwood City limits.

Phase V is the Great San Mateo / Burlingame Grade Separation.   This is a tougher project because it involves some of the most highly constricted portions of the corridor.  It also involves political and technical coordination between two neighboring cities, adding an additional challenge.  The sheer quantity of crossings (13 grade crossings + 4 obsolete grade separations within 2.4 miles) is also a complicating factor.

Phase VI is the Great PAMPA Grade Separation.  This project is left for last because it lies in the most expensive real estate on the corridor, involves coordination between three different cities, and is liable to cause the fiercest political and legal backlash anywhere on the peninsula.  Delaying it until last, perhaps into the late 2030s, allows the customarily long planning process to run its course without undue haste in all three affected communities.

We didn't arrive at today's state of grade separation (more than half) overnight.  It resulted from a slow and steady process that began in earnest in the 1940s.  The future is likely to be similar, and the peninsula rail corridor can reach a far improved state by separating just 12 more crossings over the next couple of decades, as described in Phases II and III.  This dozen should be prioritized for construction, before any of the crossings in the remaining dense clusters are touched.


  1. After recently reading your SFFS vs. FSSF post, as well as the one on getting through the San Mateo Narrows, I had a thought: When we finally get to Phase V, if the final design is a 2x2 stacked tunnel, it'd be a perfect opportunity to switch between the SFFS layout which is best for freight north of the tunnel, and the FSSF layout which is better overall south of the tunnel. If it's going to be tight and expensive anyway, might as well get the best possible use out of it, right? =)

  2. 1. What freight?

    1a. Even imaginary freight would be time-separated (even *Caltrain* gets this!), so the track layout if irrelevant.

    2. With a timetable and a service plan (see zillions of earlier posts here), there's no need to tunnels of flyovers, because there's no need for corridor-wide quad-tracking.

    3. "Tight and expensive". Yeah, I'm getting that feeling too.

    1. Uh, "anaonymous" Richard, time-separation doesn't solve the problem of nasty track-ruining 33-ton axle loads crossing onto the damn sidings. if you lay your tracks out FSSF, then the freight trains have to cross the F tracks to get to the sidings outside the Caltrain right-of-way -- over the "Fast" tracks, where you really don't want behemoth freight trains damaging the tracks.

    2. Not me.

      And you're technically wrong -- which goes without saying.

    3. Oh? My apologies, then. And since you think your brain is so much bigger than everyone else's, why don't _you- explain how 30+-tonne-axle-load freight trains, in an FSSF track layout, cross from the "S" to customer sidings outside the "F", without crossing the "F" tracks?

    4. Dear Jonathan,

      In which alternate reality does 100% segregation of "F" and "S" tracks and 100% segregation of "F" and "S" trains exist?
      In which one do 50 miles of segregated "F" and "S" tracks extend from San José to San Francisco?

      Certainly not in Caltrain's nor even in PBQD/HNTB/CHSRA's bat-shit-insane version of "reality". ("Blended plan". Use teh googles.)

      Leaving aside the objective fact that there won't be any freight traffic, let alone any worth spending tens of billions to create a peninsula-long freight-dedicated "S"-track separated from everything else. ("Port of Oakland". Use teh googles.)


  4. Clem: Your latest post illuminating the past, present, and probable future course of S. F. Peninsula Rail grade separation efforts provides a rational basis for a constructive discussion concerning the optimal course for completing the SF to SJ grade separation project. Here are several reasons for completing the SF Peninsula grade separation project with deliberate speed:
    1) Safety. A recent study showed that since public agencies took possession of Southern Pacific peninsula commuter operations over three times as many people died encountering Caltrain runs (200) as the number who lost their lives from being hit by fully grade separated BART trains (60).
    2) A completely grade separated right-of-way opens the possibility of constructing a system where no paid employee would be required to be on board an operating train. One attendant at FSSF center platforms in order to resolve far collection and train door closing issues during potentially troublesome periods, when a nearby high school lets out for example, will enable exceptionally fast and frequent 20/24 seven day a week regional transit service to be economically provided. For example 3 all stop plus 6 expresses skipping one-half of intermediate stations per hour could provide 10 minute headways and double average speeds for most passengers compared to Caltrain’s current all-stop train schedules. Note: If single deck single car EMUs only were operated the 9 train per hour total EMU rolling stock mass moved along the line per hour, total mass being a largely irreducible cost-driver, would be similar to once an hour diesel train total mass.
    3) A foreseeable grade separation completion date would enable practical track-way designs which minimize the transmission of rail operation noise to adjacent neighborhoods, produce little on-going passenger service interference while being constructed, and which eventually encourage a great deal of entrepreneurial investment in station-integrated-transit-oriented-developments. A properly executed shallow open-cut, 1,500 VDC third-rail electrification scheme could easily achieve the cost, environmental, and transit-oriented-development goals mentioned above.
    4) Historically low Long-term-bond-interest-rates plus an exceptionally competitive construction bidding environment make now an excellent time to quickly produce a Caltrain/CHSR design that is responsive to all major concerns in order to speed-up the construction planning and bidding process in order start construction soon.
    The interests of those wishing to develop a profitable fast frequent highly competitive with automobiles and airplanes peninsula rail service and PAMPA Nimbys are in fact aligned. An un-obtrusive railway grade-separation-design will encourage entrepreneurs to invest in station-integrated-transit-oriented-developments and simultaneously reduce noise transmission plus avoid visual blight to adjacent neighborhoods. A below-grade railway track resting on land-forms separated from adjacent building supports could host as quiet as possible rail operations. For example consider a pair of high-rise buildings whose foundations rest on bedrock adjacent to on both sides a below-grade track structure supported by broken rock and clay. The most economical and attractive approach in order to further prevent transmission of rail-operation noise to these adjacent buildings would be to construct an at-grade-plaza over the tracks.

  5. I don't see why you don't throw in Churchill with Meadow and Charleston.

    And I assume you've classified them as "split grade" (road goes downwards, with added maintenance, flooding, property takes, pedestrian hostility, etc) rather than "train goes up" because you're speculating what an incompetent useless transit agency will do, rather than what any body with a brain would do., right?

    1. Churchill is more constrained and in a more expensive town (north Palo Alto).

      Not everyone shares your appreciation of fully elevated viaducts... Or Berlin Walls for that matter. Churchill is *the* crossing where the whole Berlin Wall thing was allowed to spin out of control. It is best left alone for as long as possible.

    2. Well we disagree, somewhat. I'm not going to get into a blog comment back and forth, so this is it from me on this crossing:

      * Given that *the* short-term NIMBY trade-off -- ignoring the *real* issues of corridor permeability, pedestrian orientation, etc -- is between property/easement takes and large-scale road intersection reconfigurations (for split-grade sunken streets) vs Berlin Wall with no negative street changes (for rails-go-over), I think we can probably agree that the Peninsula Rail Program not only lost the plot, but never had the plot ... and has undoubtedly set back rail service on the Peninsula by yet another decade or two. (Excellent work, America's Finest!)

      * While the PCJPB owned ROW at Churchill is narrower than West Meadow and Charleston, the reality of construction phasing on an operating rail corridor is that the city street ROW of Alma will be temporarily encroached in all cases, so it's a bit academic.

      * The logistics -- road configuration, train timetable, train interlockings and signalling, simple geographical proximity, common City of Palo Alto political jurisdiction -- are so close for all three of these crossings that there would be significant design, mobilization, disruption, traffic, train operations, political savings in doing them as a package.

      That's all from me. Contradict away!

    3. Are we even talking about the same crossing? Churchill is two miles away from Meadow.

    4. I guess Richard wants a single miles-long "Berlin Wall"?

      He does seem to operate by Dogma -- as perceived by Richard, and subject to change without notice.
      Just look at his (paraprhase) "Trains go up, people and bikes stay level, Cars are Evil, and anyone who disagrees is sub-moronic". Street layout, local preference, width of ROW.. none of it makes any difference to Richard's dogmatic view how how Everything Must Be Done. It's a Diktat.

    5. The present grade-separation schemes from Churchill to Castro call for underpasses next to intersections with a fast heavily traveled Alma/Central Expressway. A bicyclist emerging from and underpass and running a light across fast traffic would likely be killed. Most drivers would avoid such an impending collision if they could readily see traffic approaching Alma on a bridge crossing a below-grade railway.
      The initial Caltrain/CHSR proposal to grade-separate the Churchill Avenue crossing with four tracks on an elevated bridge is a design that will do the most to broadcast fast train noise throughout the neighborhood thus reducing the tolerable speed trains will ultimately be allowed to run through the area. That scheme is also likely to require cutting every tree on or near the present right-of-way, disrupt on-going rail service during construction for extended periods, and end up being amazingly expensive.
      An alternative approach that might garner strong support instead of determined opposition from Palo Alto residents would be to close the Churchill Avenue railroad and Alma Street crossing to motor-vehicles, depress the railway into an open-cut beneath the current Churchill Street crossing, and construct a near-grade level pedestrian/bicycle bridge over both Alma and the railway. This bridge would provide a convenient connection to a largely car-free bicycle path from Palo Alto High School to the Bryant Street Bicycle Boulevard.
      Given the SF Peninsula’s reliably dry summer weather concrete over-crossing bridge supports and retaining walls could be constructed near the outer-edge of the track-way in slit trenches without blocking continuing rail service on a parallel grade-level track-structure. Then these future support structures could be back-filled with earth and left undisturbed for at least a year so that normal 79 mph rail service could safely resume. A Portland Cement Association sponsored chart shows concrete structures will gain 30 % in strength if left to cure for a year instead of a month in a warm moist environment; conditions easily provided for concrete structures buried in the SF Peninsula area. A 30 % lower weight and cost set of bridge pier and retaining walls would be sufficient compared to a system design where a delayed completion would prolong slow train operation.
      Subsequently during a summer weekend, when a bus bridge could connect rail shuttles on both ends of one mile construction segments, the old track-structure plus earth between the new retaining walls could be removed and new track built. The bottom of an open-cut is likely to be quite stable given that most well below grade soil has been settling undisturbed for thousands of years. This likely firm and extremely low setting rate base would enable rapid initial track construction and easily maintained smooth running conditions for high-speed-trains.

    6. There aren't any "present grade separation schemes" other than for the Rengstorff crossing, which was studied at length over several years before HSR poisened the well. Everything else is just conjecture at this point.

      Churchill carries ~10k motor vehicles per day, Alma ~14k. Not easy to close without serious impact to other parts of the road network in Palo Alto.

      Your scenario for building a trench ignores the very high water table, which would flood such a structure if it wasn't (a) fully enclosed and water-proof, and (b) pumped to remove whatever does seep in. The bottom of an open cut would quickly turn into a muddy, flooded quagmire. You need only look at what happens to the Oregon Expressway underpass during a heavy rain storm to get this point: it fills like a bathtub when the pumps can't keep up.

    7. The Oregon Expressway below-the-railway bath tub: A Google Earth altitude check shows that the entire California Avenue Business street network slopes toward the Oregon Expressway underpass. Even the California Avenue station platform has a 10 foot altitude difference slope toward the Oregon Expressway between its north and south end. The landscape surrounding the Oregon Expressway underpass is 23 feet above MSL Churchill Avenue east of the railway crossing is 45 feet above MSL; west of the railway crossing Churchill is 46 feet above MSL. A level rail connection between the Oregon Expressway overcrossing and a Churchill open-cut will stay dry with very little pumping.
      Note: A long open-cut alignment should be low enough to permit floodwater to flow on any over-crossing road bridge past the open-cut rather piling-up on retaining walls and flooding up-slope properties. As to how to design minimal cost open-cut grade separation electrified rail systems I will explore later.
      Clem: To clarify: One should prevent Churchill Avenue motor vehicle traffic turning on, off, or across Alma Street at grade. Under these restrictions Alma traffic will flow un-impeded past the Churchill Avenue alignment. A northbound Alma Street exit ramp to a westbound Embarcadero entry combined with a U-turn permitted at the left-turn at El Camino traffic light would provide a reasonably convenient near the Embarcadero/El Camino intersection PAHS student drop-off point for Northbound Alma Street drivers. The objective here is to extend a quiet easily accessible biking running and walking environment in Palo Alto and not to be excessively concerned with the number of cars being deflected.

    8. California Avenue station's existing platform falls at 0.164% north to south. It is 196.6m (645 feet) in length. You're off by an order of magnitude.

      "Un-impeded" expressway-style traffic flow on Alma Street isn't what anybody (except some troglodytes in the bowels of VTA) wants.

      If I had anything to say, California Avenue station would be elevated, California Avenue would continue level under the tracks (OH NOES BERLIN WAYY), intersect at grade with Alma Street and reconnect with severed North California Avenue, and the blighting sump pit of Oregon Expressway filled in. The gently falling grade (~ 0.4%) to the northern end of California Avenue Station would make this a rather simple and nice, also combining (grade-wise, and of course sub-urbanistically) nicely with the extra track elevation coming off a Churchill Avenue rail over-crossing.

    9. Richard: Measuring with Google Earth the California Station altitude starting at the western platform’s north end is a constant 33 feet MSL for 305 feet to the south. The remaining 340 feet to the south on the western platform descends at a nearly constant rate to 23 feet MSL when approaching the south end of the platform. (The track-way and parking lot which are nearly level with, adjacent to, and on opposite sides of the western California Avenue platform provide a broad-flat-area; a condition most likely to enable reproducible and probably accurate Google Earth altitude measurements.)
      The Amsterdam Bijlmer Arena Elevated Station you pointed to in your May 29, 2009 comment on the Shape of Palo Alto certainly looks attractive and was probably the best solution considering the probable low elevation, high water table, and responsive to the general community concern for flooding given their soggy history. Retaining the railway-at-the-upper-level grade-separation approach usually employed during the coal-fired steam engine era, when the usual means to reduce local environmental degradation was to build a tall smoke stack, makes sense if the adjacent portions of the line connected to the station being constructed are elevated.
      On the other hand when future long stretches of new urban grade separations are being considered, where air pollution due to railway operation is zero but fast train noise generates an onerous speed constraint, tracks placed in below-grade open-cuts should be considered. Quoting this forum’s Noise Calculator section: “Trenches are extremely effective at shielding train noise; they are overkill as far as noise mitigation is concerned. With a trench, even very high levels of traffic (several hundred trains a day) would create less noise than Caltrain does today. The trench need not even be covered for this benefit to be realized”.
      The ability to provide fast convenient service is essential for effectively meeting strong automobile and airplane competition. To that end it important to develop an environmentally benign rail system so that adjacent to the rail right-of-way residents will not object to fast train operation and have a low-enough above-track clearance requirement so that trains can affordably be routed to locations within a short walk to the most popular destinations. (The CPUC November 12, 1963 Amendment S-1040 to General Order 26-D allows a 12.5 foot minimum above track clearance for subways.) A January 5, 2013 Economist article on world-wide subway development indicates than an already high rate of extension of that technology should continue to grow at a 6 to 8 percent rate for the next few years compared to a 3 % growth rate for all railway infrastructure.
      In Silicon Valley Caltrain’s 12 % one-year ridership growth rate and the area’s ability to attract 41 % of the entire U. S. venture capital investment during 2011 and 2012 suggest that SF Peninsula transit-oriented development has a bright future. This opportunity appears especially strong along a rerouted Caltrain placed in an open-cut on the present Central Expressway right-of-way within Sunnyvale and Santa Clara. I suspect most real-estate developers would prefer to emphasize their own creations rather than share attention with a tall rail infrastructure looming over their new transit-oriented-developments.
      High level rail grade-separations can have an attractive appearance but like the Wells Fargo Stage optimal transportation systems in our present era are likely to have a different form.

    10. "Measuring with Google Earth the California Station altitude starting at the western platform’s north end is a constant 33 feet MSL for 305 feet to the south. The remaining 340 feet to the south on the western platform descends at a nearly constant rate to 23 feet MSL when approaching the south end of the platform."

      GE tells you the Calif Ave platform averages 1.5% descending southward? Goes to show you can't trust GE that much. Can you stop by there sometime with a level? As Richard indicated, the actual grade along the platform is probably around a tenth of that.

  6. I suggest that downtown Redwood City (Whipple Brewster Broadway Maple Main Chestnut) is much simpler than some might believe, and has higher benefits than nearly any other comparable stretch. I wouldn't throw it into the post-2050 "too hard" basket myself ... but then I'm not a World Class Transportation Planning Professional on the Caltrain gravy train.

    This sub-moronic development certainly doesn't help, though. America's Finest Planning Professionals, scoring the only goals they can: own goals.

  7. It's possible that the Burlingame—Northern San Mateo Clusterf***k could be digested in two pieces:

    * Oak Grove, Burlingame (restored, replacing North and South Lanes), Howard, Bayswater, Peninsula.

    I'd be inclined to bundle this with Broadway, but they could be done separately.
    Or Broadway plus Oak Grove could be separate from Burlingame Avenue—Peninsula Avenue.

    Elevating both Burlingame stations in a single project seems the way to go.

    * Villa Terrace, Grand Boulevard (restored).

    Could be attached to the Burlingame piece or to the San Mateo piece, more likely the southern.

    * (Villa?, Grand?,) Bellevue, Poplar, Santa Inez, Monte Diablo, Tilton, 1st, 2nd, ... 9th.

    Big fun.

  8. For Sunnyvale Ave: what would grade-separation entail? The diagrams I found on the CHSRA website show this station remaining at-grade. It isn't clear how Sunnyvale Ave would get past the tracks (assuming the roadway doesn't get closed).

    Sunnyvale Ave is a rather important crossing for bikes/peds. The only other way across the tracks is Mathilda interchange (which was designed by a crazy person). I suppose it is theoretically possible to build Sunnyvale Ave over or under the tracks, but ideally they would demolish Mathilda and start all over.

    Then again, in the ideal world Phase III would be prioritized last and tra

    1. Uggh, can't edit. Ignore that last mangled sentence.

    2. Mister Engineer,

      That's an easy one. Seriously. Unambiguous and straightforward!

      * Knock down and bulldoze all of the bat shit 1960s Mathilda elevated roadway over Evelyn and the tracks, from Washington to California. Put it back at grade, where roads that serve humans belong. This is an obvious wrong that isn't hard to right.

      (It's a nice touch -- and 100% typical of Santa Clara County, the VTA and the delightful City of Sunnyvale -- that they just spent millions of your tax dollars reconstructing this abortion. It's like a lovingly crafted historical replica theme park demonstrating Everything That Is Wrong.)

      * Trains go up and over Sunnyvale and Mathilda. Fairly easy to construct. Nice elevated station with direct bus/kiss-and-ride connections to Sunnyvale Avenue and other cut-through station access road.

      * Second stage of of phasing (after un-bit-shit-insaning Mathilda) is to knock down the Manny Valerio Memorial Parking Catastrophe (built with the full knowledge it would impede Caltrain service, and approved by America's FInest Transportation Planning Professionals still employed by the PCJPB). As usual, invite responsible parties to witness the demolition at first hand, from inside.

      * Then slew the old tracks, erect the support columns, construct the elevated track and level boarding platforms, cut over to new tracks. The usual. None of which Caltrain and HSRA's World Class Professionals will even consider.

      You're welcome!

  9. It's a nice touch -- and 100% typical of Santa Clara County, the VTA and the delightful City of Sunnyvale -- that they just spent millions of your tax dollars reconstructing this abortion.

    Richard: you failed to mention the best part of that saga. The reconstruction coincided exactly with the 2-year shutdown of Caltrain weekend service. That would have been the perfect time to have fixed that mess.

    There is also quite a bit of land lost to the "freeway" ramps leading up to the Mathilda overcrossing. If nothing else, the city could have sold those parcels off for millions. Instead, they sunk RDA dollars into that nearby 1950's mall.

  10. Good long term thoughts. Here's a great detailed thought.

    When scoping and design hearings come, or for lobbying before, the issue of roadway design speed is critical. Specs, where all this crap gets codified, for roadway design speeds can result in much larger and more intrusive projects. Why? Because roadway design speed will determine the slope and vertical curves of the roadway. A higher design speed means a flatter roadway crossing, should the roadway vertical alignment be altered. Some narrow residential streets are signed for 25 or even 15mph. Setting this as the design speed for the grade separation can allow some very tight/cosy crossings. Higher speeds get you something more like the typical Santa Clara County grade sep.

    1. The clearance for which the grade crossing is designed can also dramatically affect the size, and for now, it seems like all the road-over-rail crossings are designed for doublestack container trains, and rail-over-road crossings are designed for a full 17 feet or whatever for the road, where 13 feet would be pefectly adequate for cars, buses, and firetucks

  11. It can be easily shown that for a 1,500 VDC third-rail electrification system construction cost will be one-half or less the cost of 25 KVAC electrification capital expense for every one of the following major cost drivers:
    1) Distribution line conductive material.
    2) Distribution line suspension system structures.
    3) Total electrification system transformer costs.

    Today third-rails for new railroad electrification projects are made of aluminum with a thin stainless-steel-surface-strip for contacting rolling-stock-current-pick-up-shoes; overhead contact wire remains copper. Here is a conductive material cost comparison between 25 KVAC OCW and 1,500 VDC third-rail electrification:
    1) From Bernard de Fontgalland’s ‘The World Railway System electric traction section “The cross-section size of contact wires of ‘copper equivalent’ changes from 400 square mm for 1,500 VDC to 150 square mm for single phase AC" or 150/400 = 0.375.
    2) Aluminum does have an unfavorable conductivity compared to copper according to the following ratio: 168/282 = 0.596.
    3) Aluminum’s lower density advantage over copper is 8.98/2.70 = 3.33.
    4) The price per pound advantage of aluminum over copper is 3.688/0.943 = 3.913.
    Thus wholesale aluminum cost advantage over copper is the result of the following product:
    (150/400)*(168/282)*(8.98/2.70)*(3.688/0.943) = 2.906
    Therefore aluminum third rail LVDC conductive material costs (1/2.906)*(100%) = 34% of HVAC overhead copper wire costs.

    Distribution line suspension system structure expense comparison:
    Compared to the weight of a 25 KVAC copper overhead contact wire (OCW) weight per unit length while aluminum third weight will rise due to greater third-cross-section {(400/150) = 2.67} but falls sharply when considering aluminum’s low density {(2.70/8.98) = 0.300}. The net third rail/OCS conductive material weight comparison is: (2.67)*(0.300) = 0.802 or in practical terms nearly the same. Assuming freight operations are excluded catenary wire distributing 25 KVAC power must be at least 15 feet above the top-of-rail level. A third rail is supported every ten feet from extended cross-ties by roughly 25 cm high insulators above and outside the nearest running rail.
    Given that present rail electrification schemes either distribute DC power at or below 1,500 volts directly to EMU traction motor drivers along the line or DC power at a similar voltage within rolling stock which is generated within train-mounted transformers driven by 25 KVAC OCS distribution lines. The amount of energy per second a transformer can transmit proportional to the cross-sectional area of its electrical windings times the magnetic field area those windings enclose. Thus the magnitude of a transformer’s energy conversion rate is proportional to the fourth power of any single physical dimension. Any transformer’s weight and therefore its cost is proportional the cube of any single physical dimension. Therefore a transformer’s cost is proportional to the ¾ exponential power of its energy transmission rate. The more fragmented the power distribution system’s last step-down voltage stage requiring a transformer the higher the system’s total transformer weight and therefore construction costs are going to be. For example consider the case where 16 EMUs are supplied by a 25 KVAC power distribution line requiring a voltage-step-down-transformer on each car. What would be the total electrification system transformer weight if those 16 individual transformers were removed and those 16 EMUs were instead supplied 1,500 VDC power from one line-side sub-station transformer? Answer: 16^¾ or 8 times the weight of a single EMU transformer weight. Therefore you are cutting system transformer weight and therefore transformer cost in half by using a single line-side sub-station transformer to supply 16 EMUs.

  12. This is nonsense, because a 1500 volt DC system over a long distance requires a hell of a lot more transformers (over ten times as many) than a 25 kV AC system.

    Obviously if you have a short line and run a gazillion trains on it, the DC system is more efficient, which is why DC is used for subways. But for intercity lines, the exact opposite is true.

    Cost is dominated by the number of transformers and inverters. The international switch from 25 kV to "25-0-25" was done to cut the number of transformers in half.

    1. (Replying to John Bacon)

    2. Nathaniel: “This is nonsense, because a 1500 volt DC system over a long distance requires a hell of a lot more transformers (over ten times as many) than a 25 kV AC system.
      Obviously if you have a short line and run a gazillion trains on it, the DC system is more efficient, which is why DC is used for subways. But for intercity lines, the exact opposite is true.

      Cost is dominated by the number of transformers and inverters. The international switch from 25 kV to "25-0-25" was done to cut the number of transformers in half.”
      The complete build-out of the Caltrain/CHSR San Jose to San Francisco rail corridor envisions a complete grade-separation with passing tracks in both directions at all stations. Such a corridor has three salient properties:
      1) The number of trains-per-hour a station passing track can accommodate is usually several times the capacity of a single track at an all-stop-station. This fact enables scheduling closely following trains with varying non-interfering stopping patterns.
      2) Complete grade separation opens the possibility that a frequent 7/20 automatic train operation, where no paid employee is required for every train, could become an affordable option. For example nine EMUs per hour could then be run by a similar number of operating employees as are now required to attend one diesel Caltrain run per hour. The total weight of nine single-deck single car EMUs, a major largely irreducible cost driver, would be nearly equal to weight of one 4 or 5 car diesel train. Assigning 3 all-stop plus 6 express EMUs to call on half the intermediate stations, only the busiest of course, could double the present average en-route travel speed for the majority of passengers. Such a schedule could provide 10 minute headway service for most riders for no more than the operating cost of present hourly local service.
      3) A train distribution pattern that would most cost effectively elicit a high service quality to SF Peninsula patrons from a passing track at all stations right-of-way would consist of several long express trains stopping at heavily used stations for each short all-stop train. Since EMU traction power consumption while sustaining a 200 kmph constant speed is one-third their maximum current draw during acceleration 1,500 VDC substations should be centered at stops for most trains such as Millbrae, Palo Alto, Mountain View, San Jose, and along the maximum acceleration segments when leaving Downtown San Francisco. Cross-linked third rail conductors especially along three-and-four-track-segments would mitigate distant-from-sub-station high-current 1,500 VDC operation voltage drops. Power consumption rates would remain at a moderate level where stations are far apart or only served by short-all-stop-trains. Therefore most line-side-sub-stations should be placed next to the limited extent long-express-train acceleration paths.
      In response to the tripling of copper prices compared to aluminum beginning in 2003 a designer might choose a greater third-rail cross-section, which if following contemporary practice will be aluminum, would enable him to reduce the number of but increase the power from each substation transformer. Whether a designer uses the example I initially proposed drawn from late 20th century practice or consolidates the number of substations in response to a relative shift in the copper/aluminum price ratio the degree of power system transformer fragmentation and therefore weight and cost is far greater when comparing either 1,500 VDC case with a 25 KVAC distribution system where 160 separate transformers, one for each car, are required.
      The present Caltrain electrification proposal to permanently link four-double-deck-EMUs would reduce the on-board-transformer-weight from 10% to 7% of total EMU weight. But preventing a sharp reduction evening train length during late evening periods when demand is low but passenger sensitivity to sparse service is extremely high inhibits sharp operating cost reductions during slack demand periods without hurting service quality.

    3. Oh. So you're basically proposing subway-like frequencies and subway-like train lengths for Caltrain. In that case you would want DC.

      Good luck with that; I don't think it's remotely plausible.

      (For another detail point, on many intercity trains not every car has a transformer; DC can be passed through from car to car with AC pickup and transformers located in one place, a 'power car'.)

    4. Nathaniel: I do propose a 7/20 ten minute headway Caltrain service for most users for the following reasons:
      1) Ridership increases at a rate in direct proportion to the number of trains-per-hour increases until the rate of passenger response to shortened headways tappers-off at waiting times below 10 minutes.
      2) BART SF Peninsula mid-day ridership to their ten station South-of-Market line segment appears to be exceeding 2,000 per hour in each direction on a 16 trains-per-hour schedule at a 35 mph average speed. I am proposing Caltrain run 9 mostly one or two car trains per hour along three times the distance in order to call on twice as many stations with expresses traveling at twice BART’s average speed. A TCRP Ridership Response to Service Changes study indicates that an up-scale customer base, which is surely the case on the S F Peninsula, is more responsive than any other demographic to transit service improvements.
      3) Once a 4 track, fully-grade-separated electrified railway has been created amongst a population, that has shown a strong affinity for heavy use of a mediocre rail service compared to future possibilities, it seems criminally negligent to develop only the currently proposed warmed-over version of today’s train service quality.

    5. I would like to see "service roadmap" for Caltrain electrification.

      Current: 60 min headway. LOCAL 6AM~10PM and Complicate Express/limited each 60 min headway 6~9AM, 2-6PM.

      2013~2014: 30 min headway LOCAL 6AM~10PM and Simplified Express 6-8PM 7 days a week. Additional Express weekday commute hours.
      Local train must be 2~3 car with one conductor. Express train may runs express only between SF to Palo Alto (or Red Wood City) and then local to San Jose.

      2016~2017: Increase to 20 min headway during commute hours.

      After Electrification: 15 min headway Local, Express 7 days a week. Additional express during commute hours.

      After Transbey extenstion: 10~12 min headway during commute hours

  13. Are there no plans to do any grade separation between Tamien and Gilroy? Is electrification for the south county part of the plan? I don't see much mention.

    1. No. The brilliant Caltrain/PBQD/HNTB plan is that HSR will be a 100% separate unto itself south of Santa Clara. Because "trains" are different from "trains", you know.

      Non-electrified Caltrain is to share non-electrified tracks at grade with freight from Cahill to Tamien, and will run on non-electrified non-grade-separated freight tracks from Tamien to Gilroy.

      That way they "need" a diesel fleet and "need" to be freight compatible and "need" a three level station in SJ and "need" to keep paying HNTB and PTG and PBQD and "need" CBOSS and "need" FRA oversight and "need" grade crossings and "need" to be Amtrak compatible and don't ever need to think about "level boarding" or "operating on time" or "operating with less than three salary-sucking bodies on each train".

      It's a truly outstanding brilliant plan from America's Finest Transportation Planning Professionals, combining maximum cost, maximum regulatory grief, minimal service levels, maximum costs, maximum maintenance cost, minimal return on infrastructure and rolling stock investment and maximum consultant kickbacks.

    2. Gilroy service should be separated from Caltrain and merged into ACE or Capitor corridor. VTA should be fully responsible with this service and coorinate with thier express bus service. Shorter 2~3 car with one conductor should be good enough with ridership but need more frequency to attract riders.

    3. DMUs would probably be better to handle the passenger loads between Gilroy and SJ. Even compliant DMUs...

    4. An exceptionally powerful high speed EMU shuttle for the Gilroy to San Jose service compatible with whatever CHSR fast track is constructed between those two points might be the lowest incremental cost and certainly highest performance solution. Stopping at only the first two intermediate stops north of Gilroy would make slotting a local shuttle amongst non-stop HSR trains without interference at least conceivable.

    5. "An exceptionally powerful high speed EMU shuttle for the Gilroy to San Jose service" would be an incredible waste of money for the amount of traffic the line will have. Just buy a couple compliant DMU's and you'd be golden.

    6. It would be an exceptional waste, not incredible waste. Recall that the demographics do not justify any significant service south of Blossom Hill.

    7. If the CHSR Authority actually builds a line past Gilroy to San Jose then Caltrain Gilroy Extension rolling stock designed to minimize non-stop CHSR on-the-same-track train delays should be considered. Regional service trains using high acceleration-rate trains can reduce interference with CHSR trains competing with airplane service speed. In order to retain a safe easily maintained high-speed right-of-way eliminating avoidable passing-track switches near local stations is important. Larger than minimum capital cost electric traction motors will not only cut down the delay period due to intermediate stops but also reduce motor current for a given torque due to the larger magnetic field enclosed by given motor current plus there will be a lower motor current conductor resistance; conditions that will produce a net improvement in traction system energy efficiency in spite of the greater traction motor weight. Gilroy extension EMU traction motors should be the same as CHSR EMUs in order to mesh with HSR performance and enable efficient maintenance at an established CHSR shop. A 32.52 kw/tonne, 150 mph, 1.34 meter/(sec)^2, maximum power, speed, acceleration, and braking rate train simulation using a W = 3.026E-5*(second/meters)^2 wind constant during the power limited acceleration phase will fall 1 minute plus dwell time behind a constant speed 150 mph leader.

      What is the prospective electricity cost per seat-mile for such a 150 mph shuttle? Here is data for performance of a Shinkansen Series 300 train on a level track with a rolling lift-drag ratio of (R = 0.001) has a balancing speed of (V = 296 km/hr = 82.2 meters/second) weighing (G*M = 710 metric tons) while consuming 12,000 Kw (P = 12,000,000 newton-meters/second). Solving for the wind constant (W) using the meter-kilogram-newton-second unit system with the following equation:
      P = (R*V + W*V^3)*M*G #
      W = 3.026E-5*(second/meters)^2
      At V = 150 mph = 67.056 m/sec and applying the just derived wind-constant (W): P = (R*V + W*V^3)*M*G = 6,525 kw

      Kilowatt hours per seat mile: P/(N*V) = 6,525/(1323*150) = 0.03288 kilowatt-hours/seat mile.
      The electric power cost per seat for the thirty mile trip between San Jose and Gilroy at 10 cents per kilowatt-hour will be: .03288*30*($0.10) = $0.0986 over 30 miles per seat
      Simply raising the HSR speed (V = 220 mph = 98.35 m/s with a greater power potential but the same resistance and weight will yield: P = (R*V + W*V^3)*M*G = 20,506 kw
      Therefore the kilowatt hours per seat mile at 220 mph is: P/(N*V) = 20,506 kw/(1323*220) = 0.0705 kilowatt-hours/seat mile.