04 March 2012

The Hybrid DMU, Unicorn of the Rails

The hybrid DMU is a diesel multiple-unit train with a twist: it has on-board energy storage to enable energy recovery, similar to an electric train that feeds power back into the grid when slowing down.  Just as in a hybrid automobile, this energy store helps to start the train rolling again, and reduces energy consumption in stop-and-go operations.  Peninsula cities and stakeholders, ever more astute on rail matters, perceive three important benefits in the hybrid DMU:
  1. It keeps high-speed rail out.  In discussions of Caltrain electrification and the slow beating-around-the-bush process that is leading up to the certification of the electrification EIR, high-speed rail has become the central issue. Electrification is viewed in some quarters as the camel's nose under the tent, so a renewed push is underway to ensure that all the alternatives other than electrification have been duly considered.
  2. It keeps unsightly high-voltage poles and wiring at bay, preserving views and presumably residential property values in some of the most affluent areas in the nation.
  3. It spares Caltrain, the perennially funding-starved agency, the burden of spending $785 million of scarce capital dollars to string wires over its tracks.
The hybrid DMU is viewed as a synergistic technology that solves all three issues in one fell swoop, in a classic Silicon Valley win-win-win.

There's only one little problem: it doesn't really exist.

Okay, almost.  There do exist a handful of hybrid DMUs in Japan.  These advanced technology trains, instantly knowable to any city staffer via a simple Google search (keyword hint: 'hybrid DMU'), operate in Japan.  The fleet numbers three cars on one line, and ten cars spread among four other lines.  Each car seats about 45 and tops out at 60 mph.  Their power output is less than a Chevy Tahoe hybrid's.

The Law of Diesel Trains

To understand why the hybrid DMU will never work for the peninsula corridor, look no further than the laws of physics.  What Caltrain needs is a singular focus on better service: quicker trips, more frequent stops, higher capacity, and less waiting for the next train.  That requires big, fast trains with one key quality: punchy acceleration, precisely the reason why the EMU (electric multiple unit) powered by high-voltage overhead lines was invented and perfected.  To achieve high acceleration, the laws of physics dictate high power and low weight.

A diesel train makes all its power on board and sends it to electric motors that drive the wheels; it is essentially a rolling mini power plant.  The nicest and newest rolling mini power plants used by Caltrain today generate 3,600 horsepower for a million-pound train.  In metric units, that's about 6 kW/ton, and as any Caltrain rider can attest, it doesn't exactly pin you to your seat.  The problem is that rolling mini power plants are heavy, and if you want more power, you'll need to haul around even more weight.  This is the Law of Diesel Trains, directly derived from Newton's Laws of Motion.  More weight does little to help with acceleration, so diesel basically can't scale up.

The EMU, on the other hand, doesn't schlep around a rolling mini power plant.  Its electricity is generated off-board by a real power plant, of the PG&E variety.  The electric grid, hooked up to gigawatts of generating capacity, can provide essentially limitless power to a train.  That endows a typical EMU (of the sort available off-the-shelf from many manufacturers) with a power-to-weight ratio of about 12 kW/ton, or double the giddy-up of a diesel train.  For short bursts of acceleration, high-voltage EMUs can briefly exceed their continuous power rating and draw even more power, hitting up to 18 kW/ton.  For those still keeping track, that's triple the acceleration of Caltrain.  For reference, BART cars achieve a respectable 15 kW/ton.

So what do the Japanese know about hybrid DMUs that we don't?  Most importantly, they do not claim zippy acceleration as a benefit.  The hybrid DMU does three things for them: save about 10% on fuel, cut down nitrous oxide emissions, and cut the noise of an idling train down to an electric whisper. Beyond those benefits, the 'D' in DMU makes it follow the Law of Diesel Trains.  The Japanese hybrid DMUs manage barely 5 kW/ton, probably because they haul around not just a mini power plant, but also big heavy batteries.  Here in the US, more stringent crashworthiness standards would make such trains even heavier and their performance even more anemic.

The inescapable conclusions are thus:
  • Hybrid DMUs provide only about one third of the acceleration required to enable meaningful Caltrain service improvements. A simple technical litmus test for future Caltrain rolling stock is the power-to-weight ratio, required to be at least 12 to 15 kW/ton.  Hybrid DMUs don't qualify.
  • Hybrid DMU technology has never been scaled up beyond the size of a bus.
  • Hybrid DMU technology inherently cannot be scaled up to achieve higher power-to-weight and acceleration in large (600 - 1000 passenger) configurations.
The mystical powers ascribed to hybrid DMUs by peninsula stakeholders rightfully earns them the nickname of 'Unicorn of the Rails.'  It's time for them to realize that by undermining the choice of EMU technology and promoting hybrid DMUs, they are also undermining the future of Caltrain--intentionally or not.  Caltrain can be faulted for many things, but their choice of high-voltage EMU technology as the path to modernization is unequivocally correct and technically justified, regardless of what happens with high-speed rail.


  1. So why be coy ... who, exactly, is pushing hybrid EMUs, and in what forum? It would be interesting to know and to see what they're saying/writing about them ... and to whom.

  2. These seem low... A Prius is 53 kW/ton and runs on gasoline. Why can't a train get something similar?

    1. Trains have way lower power-to-weight ratios than cars. The best acceleration profiles of EMUs (for example, 0-100 km/h in 25 seconds) are only on a par with those of the weaker mass-market cars.

    2. Adirondacker1280005 March, 2012 12:15

      Trains don't need it, steel on steel has much less rolling friction etc to deal with. Anyway if you put too much power to the wheels instead of moving the train they just spin and grind a dent into the rail.

    3. Or in some cases, grind into the tyre of the wheel... as happened to a famous preserved LNER locmotove 20-odd years ago. ("Oops!").

      Flat tyres on US freight rolling-stock are not uncommon, and they do a power of damage to well-maintained rails.

    4. "Flat tyres on US freight rolling-stock are not uncommon, and they do a power of damage to well-maintained rails."

      Just about every single Caltrain consist, just for example, will have wheel flat spots, Caltrain-owned taxpayer-purchase underfloor wheel lathe notwithstanding.

      Caltrain is a fine example of US freight rolling stock, maintained by shade tree mechanics to the lowest standards at the highest costs.


      Simply incredible.

    5. I always thought the inherent friction difference between tires on asphalt and steel on steel only expressed itself in initial (adhesion-limited) acceleration, rather than (power-limited) acceleration at medium speed. Certainly there can exist trains with power to weight ratios approaching those of average cars, though those tend to be experimental HSR sets.

    6. I can't recall ever standing, let alone keeping my footing, in a floored Prius.

  3. Hybrid DMU's will also need batteries to store the regenerated electricity. That seems like a lot of additional weight. Does anyone have any numbers on that?

    I've revised the Latin a bit:
    vnitas mvltiplex vnicornina
    sanctvs franciscvs
    sanctvs iosephvs

  4. Your physics are wrong. Let's assume your 500 ton train, with a necessary power of 18kW/ton when starting, i.e. for 60 seconds. That would require a battery that can store 500 * (18 kW) * (60 s) = 150 000 watt hours. Ultra capacitors are around 30 watt hours / Kg or so at this point, and probably will get better. This gives a battery of 5 tons, just 1% more heavy than your initial train. Note that by using dynamic breaking and using the the large capacitor, the requirements for the on-board power plant is lower, so one could probably make it lighter.

    Trains like this may not exist, and it may be a while before commercial trains like this are available. And it may also be that these technologies in the context here are red herrings used for political reasons. But it's not an issue of physics.

    1. Ah, nothing like a theorist to ruin a perfectly valid argument. I referred to the laws of physics in a more practical, applied sense (dare I call it... engineering?).

      A hybrid DMU with high acceleration is no more realistic at this juncture than a flying car. The laws of physics theoretically don't prevent flying cars, and some examples of flying cars do exist, but there's a good reason that your and my garage don't have flying cars in them, and won't for a long while. If that's not an issue of physics, then I must reluctantly concede the point.

    2. Ant6n, is it possible for batteries to convert electricity (really chemical energy) into mechanical energy at this speed and efficiency? Car batteries are running into problems of limited range, since they need to be engineered to work in all temperatures and be rechargeable at a reasonable speed.

    3. is it possible for batteries to convert electricity (really chemical energy) into mechanical energy at this speed and efficiency?

      That's why he mentioned ultra capacitors, generically known as electric double-layer capacitors (EDLC), which are an alternate form of energy storage which allow efficient high rate charge/discharge cycles. They aren't used in used in electric or hybrid cars, as compared to electrochemical batteries they are heavier per unit of energy stored and also have a higher self-discharge rate. But, they are used in kinetic energy recovery devices in hybrid race cars (some KERS devices in current F1 cars) and buses. Oddly enough, they would be ideal for recovering energy from dynamic braking, then adding power for acceleration. Flywheel energy storage devices could also be used in a similar fashion. They may not be much of a market for relatively low power, but fast accelerating hybrid DMUs, but the technology is certainly here (though still a bit expensive) to build one...

    4. "the technology is certainly here"
      But of course!

      The future is here!

      RIght here in Silicon Valley.

      That's why little ol' Caltrain, with its world beating materials scientists, high power semiconductor expertise, unparalleled procurement skills, best in class maintenance services, unrivaled systems integration specialists, and crack operating division really needs to the global pioneer in deploying this obvious, available, cheap, simple technology! Just like CBOSS!

      Exactly like CBOSS.

      I'm only surprised that nobody has mentioned on-board thorium breeder reactors. Cheap! Obvious! The Future Without A Doubt! We have the technology! Just around the corner, for sure! A dead set certainty! Stop obstructing the march of progress!

    5. Nuclear powered commuter trains! Since you mention it, I've always thought the old gallery cars feel (and smell) like submarines. Reminds me of touring submarine museums (e.g. the old WWII relics that get docked in rivers and opened to the public).

    6. @Alon
      What Marc said.

      I would call this technology 'emerging' rather than 'available'. But given a couple more years, this should really be an engineering reality. Most of the groundwork is basically done, it just has to be put together, and the capacitors may have to get a bit cheaper and reliable etc. (which again should just be a matter of time, if the development of other battery tech is any indication).

      I think Clem's claim, especially with the comparison to flying cars, sounds like somebody making fun of electric cars in the late 90ies as something that's _inherently_ impossible because they'd either have short range, be ultra heavy, or not powerful at all. This seems strange, because usually this blog seems very factual-based and factual-like; but this particular point is a bit more rant-like than necessary. As I said, I find most points valid, i.e. that it's not a engineering reality today, and the political calculation hidden in such proposals, but it's not an engineering (/physics) impossibility. It would've been stupid in the late 90ies for a large government agency to rely on the existence of electric cars; but it doesn't make the claim that electric cars are impossible any more valid.

      Regarding CBOSS - I don't think that system is stupid because tech firms in Silicon Valley are incompetent. Rather, it's government incompetence. It's also a very specific tech, compared to ultracapacitors, which will most likely pop up in all sorts of quick charge/recharge scenarios.

    7. Adirondacker1280007 March, 2012 20:37

      I've been hearing about how, in just a few short years, I'll be fast charging the super capacitors in my car... for decades. From electricity generated in a low cost fusion plant... To drive down to the mag lev station where I'll be whisked to Washington DC in an hour and half... But then if they would just build PRT between here and the mag lev station I wouldn't have to worry about super capacitors or fusion...

    8. I've been hearing about how, in just a few short years, I'll be fast charging the super capacitors in my car... for decades.

      First off, for reason I and other have mentioned, given current technology these capacitors aren't suitable for replacing batteries in cars. Among other things, they gradually lose their charge over time. Park the car for a few hours without plugging it in, and you won't be going very far. What they are suitable for is efficient short term high rate charge/discharge, much as is needed in a bus or train that makes frequent stops.

      But, whether or not these sorts of hybrid DMUs are ultimately feasible (and with proper tradeoffs I suspect they will be), this is now meaningless in the context of Caltrain. For better or worse, Caltrain is now essentially cojoined with HSR. If HSR makes it through the peninsula, so will Caltrain. If not, my guess is that Caltrain will eventually go away, whether or not BART replaces it.

      I ride Caltrain to/from work at Stanford, and have been doing so consistently for over 5 years. I have a vested interest in Caltrain having a future. Frankly, what I've seen lately is a gradual increase in ridership, along with a gradual decline in service and reliability. Not a winning combination. In a civilized country, a commuter line such as Caltrain would have been electrified decades ago, and we wouldn't be having these discussions. All I can say, though, is just wait until there are actual plans afoot to start building out 25 KV catenary, 50 KV+ feeder lines, and the necessary substations. People will be coming out of the walls complaining about the potential of cancer and other ill effects, and others in those towns will encourage them. It's not going to be pretty...

    9. Adirondacker1280008 March, 2012 16:18

      Other than scale, what's the significant difference between applying the brakes in my car and having the need to store large amounts of energy quickly because I'm using regenerative braking and applying the brakes in a locomotive where using regenerative braking creating the need to store large amounts of energy quickly?

    10. Other than scale, what's the significant difference between applying the brakes in my car...

      I suspect scale is everything here. If I get bored later, it would be interesting to run the relatively energy requirements for starting/stopping a Prius vs. a whizzy FRA-compliant hybrid DEMU. My gut feeling is that electrochemical batteries sized to allow recovering the energy decelerating an 80 ton vehicle from 90 MPH while maintaining an acceptable charge rate would leave little room for passengers. Hence the need for flux capacitors 8^)

    11. Just like I said, the hybrid DMU is the unicorn of the rails.

    12. There is also a second difference with a train, which is that a large proportion of the braking to a complete stop is taking place at specific locations ~ the stations. Putting catenary on the run to the station and recovering the energy for the grid is quite a bit more efficient than charging and discharging a battery. And drawing power from the catenary would allow a quicker acceleration out of station than drawing power from a diesel generator and battery combination.

      At the Caltrain frequency that would be preferred, and therefore the Caltrain vehicle fleet size, catenary all the way should be a cheaper cost to own, but there may be some threshold frequency and vehicle fleet size where catenary at stops and on board power in between may be a lower cost to own. A diesel/battery hybrid system is obviously in the mix for that on board power, given that the motors are electric in any event.

      So ten or twenty or infinite years down the track when we have all the higher frequency electric heavy and light rail corridors and trolleybus networks under construction that we need to accommodate a rail mode share at Japanese levels, we ought to check back into the hybrid DMU railbus option.

      But it'd be silly for the Caltrain corridor.

  5. Another way to think about the issue is this:
    The future (such as it is) of the entire global automotive and road freight transportation industry depends on the development and deployment of high energy density, reliable, rapid charge/discharge, high power, high cycle, affordable cost on-board energy storage. But Caltrain will show them!

    Or even:

    The future of a what must be a huge percentage of the global electrical supply network depends on the development and deployment of high energy density, ... energy storage. But Caltrain will show them!


  6. Baby unicorn sighting:


    Who knows whether it could scale, but interesting technology nonetheless. And, a goat/unicorn hybrid on our favorite railroad:


    1. It's a flywheel powered people-mover, for crying out loud.
      Max speed: 65 km/hr -- 40 mi/hr. Just what Caltrain needs to get shorter trip times, hah!

  7. Clicking through on the first link to a Japanese example, there's something as striking as the short consist ~ the train running along a tree-lined hillside out in the country somewhere.

    There are not urban trains, they are what Australians would call "regional" services. When the opening of the anime Hanasaku Irohu contrasts taking the train to school in Tokyo with taking the train to get to school from a hot springs spa resort town in the countryside ... this is a modernized version of that quaint single car countryside train.

    Indeed, in the anime opening, you'll see that single car train is running on a single track line with the platforms as a passing loop.

    The engineering task of replacing a one or two car DMU on a line where a one or two car DMU makes quite a bit of sense is a quite different problem than the engineering task of replacing diesel traction on a transport corridor with the stop spacing and patronage where EMU's make far more sense. If a short DMU is already sensible way of doing things, then an incremental upgrade to that is going to be even better. If existing diesel loco-hauled or DMU technology is far from the most sensible way of doing things, then a modest incremental improvement is not going to close that gap, let alone replace an EMU as the preferred option.

  8. http://www.fra.dot.gov/roa/press_releases/fp_DOT%20107-12.shtml

    ``U.S. Department of Transportation Secretary LaHood Awards $400,000 to Develop Hybrid Road Locomotive Technologies. U.S. Transportation Secretary Ray LaHood today announced an award of $400,000 to Norfolk Southern Railroad to help continue development of an energy saving battery-operated electric locomotive that could be charged from a charging station. The locomotive could be used as a stand-alone for yard switching operations or combined with conventional diesel-electric locomotives to create a “hybrid train.”''

    Ray LaHood goes unicorn hunting!