Hacker News Reader

Not an engineer but I used to sell generators. I believe constant rpm eng/alternator is pretty common practice. From memory and a quick Google search, 1800 rpm produces 60hz and 1500 produces 50hz in newer alternators. I'm unfamiliar with these engine/alternator combos so they likely operate slightly differently.

Most power generating units operate at a relatively fixed rpm for longevity and will ramp a little during heavy load.

Source: https://www.generatorsource.com/Generator_Frequency_Conversi...

Also pulled up an old spec sheet for a Cummins and it had the same rpm/Hz specs.

Link to PDF: https://www.google.com/url?sa=t&source=web&rct=j&url=https:/...

Edit: added link to spec sheet

Older/basic generators would run at fixed frequency to generate the correct frequency. I believe the voltage is regulated by regulating the field strength in the rotor. Modern generators have inverters that handle the output and regulates the 60Hz/110V or 50Hz/220V etc. In this case the RPM is controlled by the wattage required (with a min and max RPM). Motor -> alternator -> inverter -> load

It depends on the size.

The larger the generator, the more likely you have a constant-RPM setup.

True, but that is changing. Inverters are not very expensive anymore, provide cleaner power, and can match the frequency of the something else allowing you to put several generators in parallel and disconnected one for maintenance. They also allow running lower RPM which at lighter loads can save fuel/money. (A constant rpm generator often uses more fuel at 50% load than at 80% load!)

Inverter generators can also sell power back to the grid. To pay for the fuel you need more than the typical retail price of electric, so this is only worth it if you have a need for a backup generator anyway. However if you have this need talk to the power company, they often will give you a discount because when there are load issues (See Texas last week), and the ability to add power to the grid in this time will be something they are interested in.

Today most large generators are still constant RPM, but I don't expect that to last.

Why would a constant-RPM generator not be able to feed into a grid?

Also, you can run them in parallel without problems, as long as you keep the governors from oscillating. But proper PID tuning and sensing architecture should turn that into a non-issue.

Larger engines btw. just disable cylinders for part-load operation.

They can, and do: if the governor is good enough. Overall it is easier to use an inverter, but there are several ways to solve the problem.

Cylinder disabling has a place too.

Diesel engines for power generation will have a governor control that sets the throttle based on the load to maintain whatever rpm is required for 60 Hz (constant speed control). The load in this scenario is uncontrolled and the genset simply follows it.

A diesel locomotive's engine controls are about producing the right amount of power to match the operator's throttle control. Electrical frequency at the generator (alternator) terminals is irrelevant since that output is being rectified to DC anyways. This is great for a train, because it means you can have full power at any speed - the speed of your train and the speed of your engine are completely decoupled from each other. But if you now connect this normally variable frequency AC output directly to 60 Hz loads, you will need to figure out how to set the throttle to best maintain something close to 60 Hz and your power output will be limited.

1800 rpm 60hz / 1500rpm 50hz given a four pole alternator.

One of the reasons old stationary engines (in Europe) in the pre-governator time ran at a fixed speed of 1500 RPMs

If you buy a large engine for a generator the manufacture will have one pin on the ECU for the 50/60hz switch which just changed the RPM. Generators are not enough of a market for the manufactures to make their own engines (unless they have other divisions that also need an engine for something else) that meet emissions, but they are large enough a market overall that companies that make engines want to get in on the extra profits they can by selling to the market (they get to spread the cost of emissions development across more engines), and the 50/60hz RPM needs are only a few hours to code/test on top of all the other work they are doing anyway.

There may be minimum quantities in your engine order, though the contract will provide for spare parts which might get you a single replacement many years after the generator is manufactured (if the rest of the engine is still available)

I understand that more load reduces frequency and less load raises frequency.

Maybe something got lost in translation and "specific, constant RPM" was describing (a feedback loop of) "watch the frequency and keep it steady" as opposed to "always use this value"?

--

Also, wow. Seriously, wow. A locomotive running flat out could power several hundred homes. Transportation is expensive, both from a relative cost perspective, and also from a resource(-wastefulness) perspective. Electric (specifically grid-connected) trains also suddenly make a bit more sense.

> Transportation is expensive, both from a relative cost perspective, and also from a resource(-wastefulness) perspective.

It really isn't. That locomotive running flat out may use a lot of energy. But it's also moving an enormous amount of freight, making the energy used per unit mass fairly small. For instance, CSX estimates it can move about 1 ton of freight 492 miles per gallon of diesel: https://www.csx.com/index.cfm/about-us/the-csx-advantage/fue...

Your trip to the store and back will probably use more energy.

Running rail relatively slowly and on the level is low-power. Put hills or other load factors (start/stop, speed up / slow down, even wind) into the mix and it can go up. Elevation most especially --- railroads will go up to a thousand kilometers out of a straight-line route to avoid elevation gains.

A passenger car should get 20-30 miles (30-50 km) per gallon, so unless your store run is several towns over (or you're in the sticks), probably less energy moving a ton 500 miles by rail. Though yes, personal autos are, relative to size and cargo capacity, profligate energy users.

What goes up must go down. Every joule you have to put in to climb, you get back in the downhill.

So in the end it mostly works out.

I think you are forgetting brakes. There is absolutely no way a train is free-rolling it down a hill. They will be on brakes the entire time - dynos and air brakes - so you are "getting back" a lot of those joules in the form of electricity and heat, not momentum.

Certainly. If this is a significant cost, then regenerative braking is installed. But it is mostly négligeable.

How large a battery are you planning on carrying?

None at all. Simply supply it to the grid. It would be perfectly predictable amounts of energy at predictable times, too.

Some high speed passenger trains do actually do this, FYI.

Although you can only supply the energy to the grid if your locomotive is connected to the grid while running, which diesels aren't.

Indeed, they aren't, because it's not economically worth it, because the energy is mostly too small to bother.

Which is my entire point.

Not only them. Almost all newer EMU's for regional/commuting here do it, too. And I don't like it because I can hear the fucking chirping of the electronics when they do that. Which is *very annoying' when I have to ride them.

Started around 20 years ago.

Look into why the US hasn't electrified its long-haul freight. There's a hell of a chicken-egg and scale issue involved.

1. Diesel-electric engines are not regenerative.

2. If you can't get it up, you won't be going down.

A 110 car unit coal train grosses 143 tons per car for a 15,750 ton gross trainset weight, excluding locos. Metric is nearly the same, I'll round to 15,000 tonne.

Lifting that 1,000 metres (3,300 feet) requires about 150 gigajoules of energy, or the energy equivalent of over 3.5 tonnes of oil. You're not going to recover that fuel going downhill.

(Rough numbers for an approximate answer, but it's a lot of energy.)

The highest elevation mainline crossing in the US was on the Union Pacific Railroad, Tennessee Pass (10,240 feet, 3,121 m), built in 1882 and retired in 1997. Today it's the San Luis & Rio Grande Railroad, La Veta Pass (9,242 feet, 2,816 metres).

https://www.railserve.com/stats_records/highest_steepest_rai...

Diesel electric locomotives don't use regeneration because it doesn't make much sense, not because it's not feasible. Locomotives that have regenerative braking actually do exist.

3.5 tons of fuel for 15 000 tons is 230g of fuel per ton, or around a tenth of a gallon per ton.

So even on the highest climb, assuming none of the energy is used on the downhill, that's 3 tenths of a gallon per fuel per ton for the climb.

For context, a car's engine could move the average car 5.4 miles on flat ground with the energy required for a locomotive to lift the same car up 3km.

As you can see, it's really not that much energy. Which is why regenerative braking isn't used.

Need to have somewhere to put that regenerated energy. Either overhead wire/third rail, or some sort of 'tender' with storage like the old steamers had for coal and some water.

Certainly, in general it's done via overhead wire or third rail and feeds energy back into the grid. It's not worth it in general when those aren't already present.

Well, my hybrid electric car gets, on a summer day, about 70 km per 10 kWh charge. So if the locomotive can produce about 24000 kWh in a day, it could power 2400 35 km round-trip commutes.

Now a freight train will run at, I don't know, 120 km/h? So, theoretically, it could run 2 880 km in a day. As we can see, that's a lot less than the combined 168 000 km 2400 hybrid cars could drive.

So if we're comparing a train pulled by a locomotive like this to single-occupant hybrid electric cars, the train would have to transport 60 people to be about equivalent in efficiency. That seems very doable, you could fit that in a single train car. Also the 1000 kW power output is the peak output, I can't imagine a locomotive would be anywhere near it's max power very often.

Did I goof in my calculations? I may well have! But based on these numbers, trains seem pretty efficient.

Freight train implies cargo transport, while you’re generally correct from my point of view, I think the more apt comparison would be freight train vs. semi

Also, looks like your car gets ~230Wh / mile, that’s pretty efficient compared to teslas, what kind of EV do you have? Are teslas just that inefficient or are you including ICE hybrid engine use in that measurement? Model 3 gets ~315 - 350 Wh / mile

It's an Opel Ampera, i.e. a Chevy Volt. That range, fully electric, is the best I've reliably gotten out of it, attainable on summer days (no heating!) driving on slower country roads. Highways and wintertime are a different story - this time of year I get maybe 45 km per charge. And currently it won't even run on battery only since it's so cold the car insists on intermittently running the engine for heat.

Where do you think electric trains get their electricity from?

From the skies! The overhead wires are antennas sucking it out of the air.

I used to work in the freight railroad industry. One instructive thing to demonstrate how efficient rails are, which I got to try once or twice: one person can move a fully loaded railcar with a pinch bar (basically a crowbar with an extra lever element) behind a wheel.

“ Electric (specifically grid-connected) trains also suddenly make a bit more sense.”

On a first glance it would appear so, but it a terrible idea in practice. At least for freight. Commuter trains are often electric (15 Hz? I forget), mostly for local air quality problems.

Train engines are incredibly efficient. Only the largest marine engines are better. These engines are large enough that they approach the thermal efficiency of the thermal plant producing electricity.

So no real gains for CO2 emissions. (Wind and solar don’t count at a first approx. since we’re talking about marginal E use. The train not connected on to the grid frees up production at a thermal plant to lower its production)

But you also loose a lot of resiliency. In extremis, a diesel Trains don’t need anything to run. In case of an ice storm you have to rebuild big infrastructure. Train can help you transport that if they’re independent.

Electric commuter trains have other major benefits. They stop and start frequently, and electric train sets can accelerate and decelerate faster than diesel. On the upcoming electric Caltrain, this will save quite a bit of time for long commutes. I believe that many electric trains can also regenerative brake. They can operate safely in long tunnels, unlike diesel. I suspect that diesel locomotives are considerably less efficient under the varying load conditions of a commuter train than they are for long freight routes.

And, as far as I know, all major high speed trains are electric.

> electric train sets can accelerate and decelerate faster than diesel

All locomotives use electric motors for traction. Diesel locomotives have a motor that turns a generator to produce electricity for the traction motors on the wheels. A diesel locomotive can stop and start just as fast as a fully electric locomotive, but commuter trains are relatively lighter than freight trains. Diesel locomotives can also use regenerative braking.

> A diesel locomotive can stop and start just as fast as a fully electric locomotive,

The Caltrain electrification plan respectfully disagrees. I am not remotely an expert, but I would guess there are two major factors at play. First, a lot of commuter electric trains are multiple units and have no locomotive. This means that more of the wheels can supply traction, which presumably increases the traction-to-weight ratio of the whole train set. Secondly, I think that efficient diesel generators tend to be most efficient at a specific output power, whereas a commuter train runs over a wide range of output powers. A grid-powered electric train set can briefly draw several MW from the grid without a major loss of efficiency.

To make a third factor a bit more explicit: there's a limit to the power you can apply to a wheel before it slips on the rail. Having more powered wheelsets means the total possible power is increased.

Commuter trains usually use multiple units, and metro trains always do -- they start and stop often, so the improved acceleration is a benefit.

Long distance trains can be multiple units, or a single locomotive: the time saving is less useful on a train that stops infrequently, so lower capital + maintenance costs win (plus passenger comfort, it's quieter inside).

Electric locomotives on freight trains are pretty common in Europe probably because they can have a much higher power output in a smaller unit.

A pair of TGV power cars, one at each end of the train, has a max power of 12MW at full acceleration, you would need about 14 of these diesel locos for the same output.

The power here is 25kV at 50Hz apart from some commuter lines. Also modern electric trains can use regenerative braking to dump the braking energy back into the grid. It’s pretty common.

14 diesel locs for one TGV trainset; nice way of showing how energy-guzzling high speed trains are compared to classic trains. It's still better than an airplane, but not by that much.

Actually it is a heck of a lot better than an airplane.

14 diesel locomotives sounds like a lot, but that is absolute maximum power, and a TGV carries a ton of people. On routes where they are very inefficient (London-Paris as an example), you're looking at around 10 times more emissions by plane (plane 122kg vs high speed train 11kg.) On routes where they are efficient, you're looking at more around 100x more emissions by plane (65kg Paris to Lyon by plane vs 0.7kg by TGV.)

Data here:

https://ressources.data.sncf.com/explore/dataset/emission-co...

You'll notice that you can cross the whole of France, 1001km Lille-Marseille, for just about 1.6 kilograms CO2 per person per trip. Try that in a plane!

Right but its pulling 12MW to accelerate you to 200MPH in a reasonable amount of time, it's not sitting at full power for the whole trip. I think they are capable of >300mph if you open it up to full power and see what happens. Also I got the number wrong a TGV duplex transports about 500 people with a max power of 8MW. A Eurostar is 750 people with max power of 12MW.

If you really want to have your mind blown, consider that one locomotive can move the equivalent of 400 trucks, and can be run in tandem to move more.

Trains are similar to solar panels... they are operationally super efficient but require a large capital investment. Where they work, they are magic.

Electric infrastructure is even more capital intensive, i would guess that if using was even marginally more efficient, it would have been implemented as railroads benefit from that efficiency. Equipment exists to accept grid power in diesel (Amtrak trains pulling into Penn station in NYC switch).

I think the main motivations for electrifying a railway are:

- flexibility and independence of power source. Switzerland can use hydropower, France can use nuclear power, many places could use coal power and not worry about losing rail transport in an oil shortage.

- reduced maintenance costs (no diesel engines; reduced weight of the trains causes less wear on the track)

- more power, i.e. faster trains, increasing the capacity of the railway line. Passengers approve, and freight trains can accelerate better meaning it's easier to run them between passenger trains.

The US and Canada have less to gain with these than many other countries.

> Also, wow. Seriously, wow. A locomotive running flat out could power several hundred homes.

That's exactly what happened in my hometown, which generated power via a locomotive that was literally up on blocks. You could hear it grinding away from quite a distance. They only got connected to the grid when I was in my 20s and long gone.

“but few loads really cared too much.“

There’s some room for error, but frequency is the a very important metric for the utility to get right. The reason is that motors start to burn when the frequency is significantly off of the design f [1].

So, the blower in your furnace will die (-> no heat in the home). Your laundry machines’ motor burn. You ovens convection oven burns. Etc.

You’re better of with rolling black ours (Texas forgot the “rolling” part) than get the frequency wrong.

[1] think of a motor as an LCR circuit. There’s a natural f, and if youre far from it it will cause heating of the windings

> Your laundry machines’ motor burn.

Modern laundry machines have sophisticated power electronics between mains and drum motor.

Operating in notch 6 runs the generator at 647 RPM. Since the AR10 is a 10 pole machine, that gives 64.7 Hz power.

This part doesn't quite make sense though, because it mixes RPM (per-minute) with Hz (per-second) without scaling.

647 RPM is 10.8 Hz, and more poles should increase the frequency not decrease it - 10 poles means you'd multiply it by 5, so would give you 54 Hz.

probably 54 Hz- split phase power has a GND, positive phase, and negative phase: https://techblog.ctgclean.com/wp-content/uploads/Split-Phase...

Alternating phases would be hooked up to one phase or the other, with negative phase poles connected backwards. I know traction motors are brushed, so probably the generators are as well. If that's the case then you should be able to just bolt new wires on (brushes are consumables, so they are easily replaceable).

edit: managed to misread your conclusion as it would be 110 Hz, lol.

I wonder what they did with the 5 phases? Use them asymmetrically and let the engine struggle with the unbalanced load? Also it seems like this feed would be no good for powering 3 phase motors, or really any phase-phase load. Although maybe city hall back then was only on single phase power, and they could balance circuits between all 5 phases, with the probable addition of temporary electric heaters etc.

With what part of it do you take issue? I don't see the discrepancy...

I think the only actual reason why the power frequency has to be precise is so all the power plants can be synchronized. But in an emergency situation, you don't really care if it's 60 or 65 Hz...

Wild variations are bad for heavy inductive loads like transformers and motors. Within 10% is unlikely to damage anything though.

In the video, the French-speaking news anchor says that one of the biggest challenge was to convert the DC from the engine to AC (around 1:10).

Slight change of interpretation: the biggest challenge was to disable the AC->DC rectification circuitry, the generator already made AC.