The plausible outlets are reporting that both 3GS21Bs and single 3GS21C have been retired and sold back to National Railway Equipment. This surprised me as I had expected the 3GS21Bs to be rebuilt into roadslugs and the 3GS21C fodder for the SD33ECO program… unless the deal is for NRE to send NS five SD40s/45s to fill out the SD33ECO program.
Evidentally they’re worth more intact than as parted out hulks that might be utilized as frame/running gear donors for yet another project.
They’re relatively young and surely would be a good fit somewhere, where as they were oddballs and apparently somewhat unsatisfactory in the service NS utilized them in. Union Pacific and BNSF roster over 100 3GS21B’s for instance, so there has to be some decent value there as an intact locomotive.
Will be interesting to see if any of the other low-emission oddballs follow shortly in being strickened. While the two RP20DB’s are active, both RP14BD’s and the single RP20CD are stored. And then there’s the PR43C fleet that is sitting idle rusting away right now and while active, battery powered BP4 doesn’t look to have been a complete success and likely isn’t the prototype for an entire line of sisters.
NS had already gone on record with their disappointment with the performance of gensets.
Having traveled far and wide across the North American rail system and having visited just about every major rail yard there is one commonality found with gensets: they were always parked.
Maybe NRE has a different use in mind for them then rather than reconditioning them for resale. Maybe they’ll part them out and use the frame, body, and running gear for a different project that aims to improve upon some of the ailments from their earlier attempts.
Either way, they evidentally were worth more via resale than they were to NS for any potential use.
Crews hate GenSets - not enough power when you want it, by the time full power comes on line - you don’t need it. A bad solution looking for a problem.
The idea’s not wrong – just the implementation.
The problem in part is that people who don’t understand switching were designing the things. Switch engines aren’t little underpowered Thomas clones that shuttle hither and yon with cuts of cars, any more than commuter locomotives are old hand-me-downs trailing a sorry bunch of Ping-Pongs on a slow trek in and back. And the dominant characteristic in both cases is the same: the native ability to accelerate very quickly with a load, and just as quickly scrub momentum off.
It is extremely unwise to use ‘conventional’ battery chemistry as the source or sink of rapidly-increasing drain or charging current. Even if you can modulate this so the battery ‘lives’, ther needed to be an explicit control that predictively fires up one or more of the engines in time for the equivalent of ‘power mode’ (best power from the effective ‘sum’ of engines and battery) when performing a kick, and winds the engine down to minimize emissions afterward. And, for that matter, reduces other forms of pollution by spooling the engines up in speed with reduced load, to cut down the nanoparticulates and some other undesired exhaust constituents that form if the engine has to accelerate into substantial load, but then haqs the ability to put engine horsepower quickly to work without lugging.
I was surprised that GE in particular did not take an interest in marketing its ‘hybrid battery’ technology to this market … or that we haven’t seen my old ‘pet’ technology for this, some big counterrotating flywheel storage. Perhaps the market isn’t exactly there. I can’t help but think, though, there’s a good market especially in air-quality management districts, for a genset-style switcher that can actually switch.
I believe that the power in question are just gensets, not hybrids like the various Green Goats from Railpower Technologies.
They’re powered strictly by diesels, with multiple engines that come online as needed to meet the demands at that time (Or at least that’s the theory).
It’s supposed to be more environmentally friendly. A NRE 3GS21B for instance can be a 700 hp, 1400 hp, or a 2100 hp switcher as needed instead of underutilizing a 2100 hp switcher in service where often only a fraction of that power is needed.
Clearly it’s not yet a perfect solution and I suspect it may never be until a traditional builder tackles the concept with a product designed and built specifically for railroad use in mind, instead of repurposing major components from elsewhere like the construction equipment engines that are in the 3GS21B. Not to mention their support infrastructure and reputation backing it.
They may be easy to remove and reinstall during maintenance, but if they’re breaking down frequently and don’t age gracefully, it matters little.
For switching service, my take is that “ultracaps” are a better match than batteries. Capacitors provide a much higher specific power, efficiency and cycle life than batteries. Downsides are that specific energy (e.g. w-hr/lb) is lower and the voltage rises and falls with the state of charge.
Maxwell Labs makes a 48V/165F capacitor module, which weighs about 30 pounds and QTY 1 price from Digi-Key is about $1350. The module is rated for 1,000,000 cycles (most batteries do well to reach 500). Assuming an operating mode where the capacitor is cycled between 50% power (70& voltage) and full charge, each cycle would be good for 25 w-hr. 1,000,000 of such cycles works out to 25,000kwhr or about $0.05/kwhr. Caveat is that you need to do 100,000 cycles per year or about 12 cycles per hour.
25 w-hr is enough energy to accelerate a 1,000 ton train from 0 to 1 MPH, 225 of these modules would be able to accelerate that same train from 0 to 15 MPH and with regenerative braking be able to recover most of that energy in stopping said train. With that power available instantly, this seems to be ideal for switching service.
- Erik
FWIW, Eaton appears to be a second source for ultracap modules.
Nobody but an idiot sources or sinks high current through a battery – especially not in repeated charge and discharge cycles, as the Green Goat people discovered (and in my not-so-humble opinion should have recognized from the outset)
There are additional issues of concern with ultracapacitors. One is that they are inherently very-low-voltage devices with very strict requirements about overvoltage (even short-term spike voltage), so substantial connections in series are needed to produce the necessary voltage, and very good power conditioning necessary (it can also be difficult to assure a good ground path for ‘clamped’ spike attenuation). There is also a concern about how the devices tolerate internal damage, or if they can be made self-healing (or self-de-shorting, to coin an ugly term) so that relatively long life and graceful degradation of capacity can be observed in practical terms. They’re much better suited to be “charge buffers” for handling high transient current than storing electrons at required energy density for a couple of cycles of kicking and braking.
One of the traditional problems with gensets is that, for pollution reasons, there are various delays associated with startup and power changes. Some of these can be addressed with better design, for example keeping all the engines on a common ‘cooling’ circuit that maintains the general block temperature (and more particularly the liners adjacent to combustion) at full operating temperature, and reducing the necessity for glowing before admitting fuel at starting. Many concerns here are also concerns for diesels that can tolerate ‘shutdown and restart’ in motor-vehicle use, for example at stoplights or when coasting, so there is ongoing research into various aspects of “pollution control”. Another concern (which is shared with more conventional large diesel engines) is the rate at which a diesel changes speed, especially when under load; I
I always thought gen-sets had an awful lot of moving parts for what they did. One 12 cylinder EMD turbo has the same output as those three engine gen-sets.
I also thought the benefits of a gen-set would be marginal. That 12 cylinder EMD’s fuel efficiency is nearly the same in any notch (once the idle fuel burden is paid for). So, maybe the gen-sets save you some idle fuel? Let you shutdown in cold weather? That’s not much… For a yard/local locomotive, I’d take the wayside plug- heater solution over a gen-set.
Maybe NS has reached the same conclusion?
You take a Vapor-Clarkson automatically controlled oil fired 300-psi steam generator and connect it to a heavily insulated water tank with a steam space on top. You feed steam to an eight coupled wheel engine set.
Besides providing good insulation blankets over everything, you provide steam jackets to the cylinders to reduce condensation losses, especially for operating after the thing is parked awhile. The steam jackets don’t have to be complicated cylinder castings. Rather, they can be pipe traces welded to the cylinders and then covered with the fiberglass or other rock wool insulating blankets.
Bam!
Tractive effort on demand and low standby losses . . .
Actually, what you do is charge the device from an external overcritical-water source, like a power or utility boiler, and use the on-board heat source mostly for for ‘makeup’ and superheating. I think there are better onboard steam-generator designs than Vapor-Clarksons (although, with modern controls and implementation, they are a reasonably evolved and efficient design) – see the Cyclone Engine for one alternative design that also generates some topping shp.
I also wouldn’t do typical reciprocating drive – I’d do something like Steins’ oil-fired switcher, the poster child for Franklin type D if the patent drawings are accurate, or a modernized Heisler arrangement with closed gearcases to ‘diesel’ wheelsets. I won’t rehash all the arguments unless someone wants to get into them. Yes, you could use modulated independent braking on driver cheek plates to do practical slip/traction control for high acceleration at low effective FA, so I’m certainly not going to agitate against the principle of relatively-simple recip engines in this service, and yes, there is some advantage in using modern materials like cerium steel in reducing augment in the running gear (“high speed” here being taken as, properly I think in this context, high rotational speed) BUT it needs to be remembered that this service is one of the worst for rapid deceleration with large trailing load, so buckling or deflection of the rods is a much more likely and severe issue than in most forms of road service – designer beware!
And that has improved, radically, in the past decade with aerogel, nanoshield, and ot
Li-ion batteris are more than capable of sourcing or sinking high currents - state of the art ten years ago were that Li-in batteries had about the same specific power as ultracaps. Cycle life for batteries is an issue though.
Part of the reason for using the Maxwell modules is that the modules take care of the voltage balancing, monitoring OV condictions and over temp. These modules are designed to be connected in series, with the proviso that the string voltage does not exceed 750V. A string of 15 of these modules in series would have an normal max voltage of 720V, with 765V being the Do Not Exceed limit. For a single series string, going from 720V to 765V would require over 20,000J of energy. My proposal was to use on the order of 15 these strings in parallel to get a useful energy storage for
[quote user=“RME”]
Paul Milenkovic
You take a Vapor-Clarkson automatically controlled oil fired 300-psi steam generator and connect it to a heavily insulated water tank with a steam space on top. You feed steam to an eight coupled wheel engine set.
Actually, what you do is charge the device from an external overcritical-water source, like a power or utility boiler, and use the on-board heat source mostly for for ‘makeup’ and superheating. I think there are better onboard steam-generator designs than Vapor-Clarksons (although, with modern controls and implementation, they are a reasonably evolved and efficient design) – see the Cyclone Engine for one alternative design that also generates some topping shp.
I also wouldn’t do typical reciprocating drive – I’d do something like Steins’ oil-fired switcher, the poster child for Franklin type D if the patent drawings are accurate, or a modernized Heisler arrangement with closed gearcases to ‘diesel’ wheelsets. I won’t rehash all the arguments unless someone wants to get into them. Yes, you could use modulated independent braking on driver cheek plates to do practical slip/traction control for high acceleration at low effective FA, so I’m certainly not going to agitate against the principle of relatively-simple recip engines in this service, and yes, there is some advantage in using modern materials like cerium steel in reducing augment in the running gear (“high speed” here being taken as, properly I think in this context, high rotational speed) BUT it needs to be remembered that this service is one of the worst for rapid deceleration with large trailing load, so buckling or deflection of the rods is a much more likely and severe issue than in most forms o
[quote user=“RME”]
BaltACD
A bad solution looking for a problem.
The idea’s not wrong – just the implementation.
The problem in part is that people who don’t understand switching were designing the things. Switch engines aren’t little underpowered Thomas clones that shuttle hither and yon with cuts of cars, any more than commuter locomotives are old hand-me-downs trailing a sorry bunch of Ping-Pongs on a slow trek in and back. And the dominant characteristic in both cases is the same: the native ability to accelerate very quickly with a load, and just as quickly scrub momentum off.
It is extremely unwise to use ‘conventional’ battery chemistry as the source or sink of rapidly-increasing drain or charging current. Even if you can modulate this so the battery ‘lives’, ther needed to be an explicit control that predictively fires up one or more of the engines in time for the equivalent of ‘power mode’ (best power from the effective ‘sum’ of engines and battery) when performing a kick, and winds the engine down to minimize emissions afterward. And, for that matter, reduces other forms of pollution by spooling the engines up in speed with reduced load, to cut down the nanoparticulates and some other undesired exhaust constituents that form if the engine has to accelerate into substantial load, but then haqs the ability to put engine horsepower quickly to work without lugging.
I was surprised that GE in particular did not take an interest in marketing its ‘hybrid battery’ technology to this market … or that we haven’t seen my old ‘pet’ technology for this, some big counterrotating flywheel storage. Perhaps the market isn’t ex
Well yeah, superheat has two functions. One, it increased the thermodynamic cycle efficiency somewhat by expanding the specific volume of steam at a given pressure. Two, it keeps everything hotter so there is less of all modes of power-robbing condensation.
I have read enough to realize that a wall jacket, whether steam or as suggest pressurized water for this application, does something that a high degree of insulation doesn’t quite provide. I have been told and I have also read of a “poor man’s” wall jacketing in the form of training crews to set the brakes, open the throttle, and cycle the reverser so as to heat the cylinders up before starting up.
The evidence from Chapelon’s 160 heavy tractive effort locomotive experiment is that if you keep the steam hot, from wall jacketing and compound-expansion reheat (a kind of superheater between HP and LP cylinders), the thermodynamic advantage of superheat is only about 7 percent compared to 30-40 percent when you rely on superheat to keep the walls hot.
So, in place of superheat, which is difficult to achieve in the sort of steam generator fed accumulator locomotive I have in mind, you could do compound expansion and then reheat the steam between HP exhaust and LP inlet? Or you could apply heat to the cylinder jacket, or do both.
“Advanced” steam cycles in stationary po
Chapelon did both in the 160A (he had a little Schmidt-type ‘resuperheater’ in the lower part of the boiler). I also recall his proposing using superheated steam injection as a means of achieving better equalization of cylinder effort between HP and LP (for example to simplify some aspects of dynamic engine balance)
Actually, what you see more often for ‘thermodynamic advantage’ are more than one bleed in between turbine stages, to optimize the feedwater heating part of the Rankine cycle. There is usually very little advantage in passing turbine flow out of the rotating machine to anything that can impart the necessary fast heat transfer with concomitantly minimal flow restriction and then pass the flow back into the turbomachinery without having a blind section of shafting in the middle…
(Now, that gets done for turbines in some PWR nuke plants (where the steam is not really ‘superheated’ but there is LOTS of cheap thermal energy for the taking. Nobody likes to talk about ‘thermal efficiency’ too much when discussing those things…). One of the first full-scale reactor setups (at Indian Point) was specified and I think built with separate oil-fired superheaters to get the steam quality where the coal-plant engineers thought it should be for best efficiency. Turned out there were other ways to optimize a fuel source sometimes …
[quote]
You can think of cy
I got lost. Are we talking about steam gensets?
We’ve gone from the science of gensets to invoking the name of Andre Chapelon. I love it.
Not quite. We’re talking about a steam locomotive that does a specific job – switching where cars are frequently kicked – better and with the same sort of economy as genset locomotives are supposed to provide.
First: As Ed and some others can tell us, a large range of practical switching jobs involve high acceleration and rapid braking. This requires – for very short periods of time – very high power, while ‘the rest of the time’ the powerplant of the switch engine is essentially idling (or shut down) when not needed.
The premise of the genset locomotive was to use multiple small engines that could be started and run in combination, to select the amount of power needed at a particular time. This was great for limiting emissions in air quality management districts, but not so great when the amount of power developed by the locomotive has to go from low to full quickly, without more than a few seconds’ actual warning, and then as quickly be reduced to ‘coasting’ level. Diesel engines don’t like accelerating into a load, and they (and their turbochargers) don’t like being chopped back to idle or, worse yet, shut off as soon as they aren’t needed.
This issue can be partly overcome by providing an energy-storage transmission – what in cars is called a hybrid – which can store energy over time from a comparatively small or lightly-loaded powerplant (running steadily at the peak of its torque curve where efficiency is highest and emissions at a minimum per hp/hr), release the stored energy quickly to aid acceleration, and then use regeneration to provide both braking and energy recapture. To this, we add the ability to fire up additional engines ‘when needed’ just to help accelerate cuts of cars a bit quicker.
What Paul was proposing is to use a fireless cooker locomotive with a compar