Regenerative braking on Locomotives with AC transmission

[rollingstein]

Many modern electric train locomotives have the option of regenerative braking. Most of these locos have HVAC power transmission.

I'm wondering, when they brake in regenerative mode do they actually generate DC (I'm not sure whether modern electric locos prefer DC motors or AC motors or is the choice still open), then invert it to the right AC frequency, then match the frequency / phase on the fly precisely to the grid & then transmit power back?

Sounds like a lot of work especially if you factor in the large power loads involved e.g. a typical electric locomotive can easily exceed 8000 hP though I'm not sure how large the max regenerative braking load can be. Anyone know?

 

[Averagesupernova]

I would think some creative google searches could turn up some info. Your suggestion is something I hadn't thought of. Last I knew all the diesel electrics dump the power generated by braking.

 

[rollingstein]

Your suggestion is something I hadn't thought of. Last I knew all the diesel electrics dump the power generated by braking.

Well, the diesels have no choice than to dump power as resistive heating because there's no transmission line they can dump to.

I think that's the difference between dynamic braking (dump your power as heat to a heavy duty, fan cooled, resistor grid on board the loco) versus regenerative braking (ship power to the grid).

Even among electric locos I think some use only dynamic braking but I guess that's more wasteful.

 

[rollingstein]

Just as a concrete example:

One of the large electric locos used in China is the HXD2B with a tractive power of ~10 MW & a max allowable regenerative braking load of almost the same amount. (Source Wikipedia)

This loco seems to have AC motors.

 

[zoki85]

AC motors, in regenerative breaking, generate AC power (async. generator mode) and feed the network.

I'm wondering, when they brake in regenerative mode do they actually generate DC (I'm not sure whether modern electric locos prefer DC motors or AC motors or is the choice still open), then invert it to the right AC frequency, then match the frequency / phase on the fly precisely to the grid & then transmit power back?

 

[rollingstein]

AC motors, in regenerative breaking, generate AC power (async. generator mode) and feed the network.

Thanks. Would this AC power be always of a fixed frequency irrespective of the speed of rotation of the decelerating, braking wheels?

I think not. If so how do they get it up to the grid frequency? Do they rectify this variable freq. AC & then re-invert it using a static inverter to the desired grid freq.?

Also, they must have to match phase too, using separate circuitry?

 

[zoki85]

AC→DC→AC conversion systems are most frequently used today.

 

[jim hardy]

They know wheel speed and apply AC at frequency to make desired torque, be it positive or negative.
http://www.republiclocomotive.com/ac_traction_vs_dc_traction.html

The AC system, however, operates in a very different fashion. The variable frequency drive creates a rotating magnetic field which spins about 1% faster than the motor is turning. Since the rotor cannot exceed the field speed, any wheel slip is minimal (less than 1%) and is quickly detected by the drive which instantly reduces load to the axle.

.........

With AC traction, it is also important to consider braking. As with traction, braking is a function of weight on drivers. Therefore, when using standard friction braking (tread brakes) the braking capability of the locomotive (excluding train braking) is proportional to the locomotive weight. With AC traction, however, the braking can be much higher because the drive system in braking acts just like the drive does in traction thus eliminating wheel slip. The drive converts the motors to generating mode (dynamic braking) and the electricity produced is dissipated in the braking resistors. Thus the motors are slowing the locomotive without using the air brakes. Again, the adhesion levels are much higher so the locomotive can again be significantly lighter for the same amount of braking. The dynamic braking in AC traction locomotives also allows full braking down to zero speed, unlike DC dynamic braking.

 

[rollingstein]

AC→DC→AC conversion systems are most frequently used today.

Thanks for clarifying that.

I was earlier confused why an AC motor loco using transmitted AC power had in its specs a rectifier & inverter system.

The AC to DC to AC bit explains that I guess.

 

[rollingstein]

They know wheel speed and apply AC at frequency to make desired torque, be it positive or negative.
http://www.republiclocomotive.com/ac_traction_vs_dc_traction.html

Very interesting article, thanks.

But that explains traction & dynamic braking, right? My core confusion is regarding regenerative braking.

In dynamic braking what frequency / voltage / phase you generate at doesn't matter much because you are burning away the power in a local resistor grid. But not so in regenerative. There if you must feed into the grid all three (frequency / voltage / phase) will have to be precisely matched.

 

[jim hardy]

Diesel electrics aren't connected to the grid .. at least the tracks near my house have no wires overhead...

I don't know about those fully electric trains with overhead wires, i suppose surely they would return power to the source.
But that's just a guess.

An AC induction motor rotating faster than the source supplying it will return power to that source. That's called "negative slip", and is a handy term for electric motor guys..
but the locomotives i watch go by here are at idle when coasting downhill. I think you'd want to not risk overspeeding the engine when coming down out of the mountains, so dump the energy as heat . There's noplace else to put it for they can't turn it back into diesel oil...
The key is with electronics they can create whatever frequency they need, per that article slightly above or below wheel speed so as to modulate power flow into or out of wheels by making either slip positive or negative.

I guess if "regenerative" is defined as making useful work out of the energy that's been re-generated by the traction motors, then it's not "regeneration".
Eternal Vigilance is the price of precise vocabulary, eh?

Interesting discussion. Thanks for letting me play !

old jim

 

[rollingstein]

Diesel electrics aren't connected to the grid .. at least the tracks near my house have no wires overhead...

Indeed. Diesel electrics cannot regeneratively brake. They can only dynamically brake i.e. burn up the EMF generated as heat from a resistor grid. (unless the diesel electric was pulling a Passenger car load and used the EMF for hotel power; not sure. ) Yes, regeneratively means using the power generated for something useful. Otherwise it is dynamic braking.

The question I had was in the context of the true electrics. The one's pulling power from a panto-graph and overhead centenary.

Interestingly the most powerful locomotives today are almost all true electric and not diesel electric. (I think)

 

[Windadct]

True Re-gen needs somewhere to put the energy - and in trains it is a lot. So this is in Electric Trains connected to AC or DC supply.

The technology is pretty common - you can also look up regenerative AC motor drives. These do require an inverter on the supply side - and during re-gen that inverter has to synchronize with the supply to be able to "push" power ( energy) back to the source.

The secondary advantage is that this configuration usually runs the inverter as an ACTIVE RECTIFIER when it is converting the AC to DC - this makes for a much improved current waveform ( better Power factor and harmonics) than a passive rectifier with just diodes. In large system this actually improves system capacity because the supply would normally have to support the bad PF ( reactive current) or Harmonics.

The Basic Topo will look like this in IGBT types

 

[rollingstein]

The technology is pretty common - you can also look up regenerative AC motor drives. These do require an inverter on the supply side - and during re-gen that inverter has to synchronize with the supply to be able to "push" power ( energy) back to the source.

Thanks @Windadct!

Now we are getting somewhere! Thanks for the details.

So, in the diagram you provided I understand the topology for the AC -to DC conversion. i.e. rectification. But following this DC link there is an inverter right? Generating a variable freq. AC to drive the motors?

Now in regen do all those blocks get used in reverse? That's the part I'm unclear about. i.e. While doing traction the grid automatically gives you a stable freq. and Voltage whereas in regen the circuitry to match phase, freq. and V will have to be explicitly provided, right?

 

[Windadct]

I think you have it - the output of this system is variable speed (to the drive motor) - it should be clear cut what is going on, however for the regen - think of a (sine) wave, you fire the IGBTs to just lead the utility 60 Hz - just making 60 Hz is not really ideal ( here in the NE of the USA there is Amtrac at 25 Hz! - must have been a steel barron that made that decision!) . Anyway - you want to be a little ahead of the curve - you then are providing reactive current and power.

So in Regen - you want higher V and very slightly leading PF ( typically! ) -Higher V is actually easy - since the Anti-parallel diodes in the inverter are now just rectifiers - the DC link voltage can get quite high as it absorbs energy from the motor

 

[rollingstein]

So in Regen - you want higher V and very slightly leading PF ( typically! ) -Higher V is actually easy - since the Anti-parallel diodes in the inverter are now just rectifiers - the DC link voltage can get quite high as it absorbs energy from the motor

Thanks!

V & freq. we took care of. But what about phase? Isn't that critical too? There ought to be no reason why the regen voltage will fundamentally by in sync. with the transmission line, correct? You'd need circuitry to explicitly do that phase matching business, right?

In the whole rectifier-inverter AC drive system normally phase matching isn't a concern at all, is it? But for regen it should be?

 

[Windadct]

Huh? ... Certainly? - If I get your post correctly, yes, on the Grid / source connection side you have to match the phase for sure, my point being that the grids V is essentially fixed - to push power back onto the grid you need higher V and generally a slightly leading PF ( once you are synchronized)

 

[rollingstein]

Huh? ... Certainly? - If I get your post correctly, yes, on the Grid / source connection side you have to match the phase for sure, my point being that the grids V is essentially fixed - to push power back onto the grid you need higher V and generally a slightly leading PF ( once you are synchronized)

No,. sorry. My bad. I should have been clearer.

I agree with what you wrote. I am only asking if due to regenerative braking you must add special circuitry to the locomotive to perform this extra task of phase matching.

Normally for traction nor dynamic braking you'd never need this phase matching circuitry (I'm thinking). If so, I'm wondering how does that system look like. With IGBTs / thyristors etc. how does one design a phase matching circuit.

 

[anorlunda]

Whoa whoa, you are all ignoring what Jim Hardy said about induction motors (see the quote below) No fancy inverters, or frequency or phase matching needed. But you do need railways with electric power lines. They are quite common in Europe and they have been inherently regenerative for a long time.

Diesel electrics aren't connected to the grid .. at least the tracks near my house have no wires overhead...

I don't know about those fully electric trains with overhead wires, i suppose surely they would return power to the source.
But that's just a guess.

An AC induction motor rotating faster than the source supplying it will return power to that source. That's called "negative slip", and is a handy term for electric motor guys..
but the locomotives i watch go by here are at idle when coasting downhill. I think you'd want to not risk overspeeding the engine when coming down out of the mountains, so dump the energy as heat . There's noplace else to put it for they can't turn it back into diesel oil...
The key is with electronics they can create whatever frequency they need, per that article slightly above or below wheel speed so as to modulate power flow into or out of wheels by making either slip positive or negative.

I guess if "regenerative" is defined as making useful work out of the energy that's been re-generated by the traction motors, then it's not "regeneration".
Eternal Vigilance is the price of precise vocabulary, eh?

Interesting discussion. Thanks for letting me play !

old jim

 

[rollingstein]

No fancy inverters, or frequency or phase matching needed. But you do need railways with electric power lines. They are quite common in Europe and they have been inherently regenerative for a long time.

Interesting.

Well, forget regen but with no inverters how do these AC induction motors achieve speed control during traction duty? Can you describe the system in more detail?

 

[anorlunda]

Interesting.

Well, forget regen but with no inverters how do these AC induction motors achieve speed control during traction duty? Can you describe the system in more detail?

That is a very broad field with many companies In many countries doing it for more than a century. I suggest that you start with wikipedia "railway electric traction" and induction motor, and traction motor. The traction motor article mentions regenerative braking.

But to return to the OP: Where railways are diesel powered with no power lines connected to the train, then feeding the grid is impossible. Where railways are electric (Wikipedia) they do not use HVAC transmission voltages. They use relatively low voltages (600 V DC to 25 KV AC) because the lines are so close to the ground. Also the same electrocution safety issues would arise as does back feeding the grid from rooftop solar panels. Finally, you must allow for the case when the grid can not take the regenerated power and provide the dynamic braking system as a backup -- double capital costs. The regenerated energy would have to be very valuable to make it interesting.

 

[rollingstein]

I think you'd want to not risk overspeeding the engine when coming down out of the mountains, so dump the energy as heat . There's noplace else to put it for they can't turn it back into diesel oil...

Ha. I found another place to dump the energy generated during regen braking. I quote from an actual diesel-electric 5000 hP loco spec. I dug up:

"Hotel load controls shall preferably be designed such that regenerated dynamic braking power is fed back to hotel load inverter. During dynamic brake, the system shall be able to use dynamic brake power to the extent possible for hotel load and any short fall shall be met from traction power."

So, one way to usefully use the regenerative braking electricity is to use it to light up the ( passenger) train behind you instead of burning it up as heat in a dynamic brake's resistor grid.

 

[jim hardy]

I quote from an actual diesel-electric 5000 hP loco spec. I dug up:

Great find ! Would you post a link to that manual?

 

[rollingstein]

Great find ! Would you post a link to that manual?

Sure. Here it is. Isn't a manual. More like an RFP for quotations.

http://www.rdso.indianrailways.gov....l-Cab-4500HP-67-Rev-03-uploaded-08-Aug-12.pdf

 

[rollingstein]

It'd be interesting to calculate if a 500 kW hotel load would be enough to dissipate all the regen. braking power generated.

Say you want to hold speed steady of a 1% down gradient (fairly typical as a max?) for a 18 coach passenger train how much of power dissipation would one expect? Using no air braking.

 

[jim hardy]

Well let's ballpark it
Amtrak Superliner cars weigh 74 tons
http://en.wikipedia.org/wiki/Superliner_(railcar)
17 of them would be 1332 tons

add maybe 200 tons for a diesel electric locomotive
though they can easily be twice that
http://en.wikipedia.org/wiki/GE_AC6000CW

and we're around 1500 tons.
40 mph is 58.7ft/sec, call it 60

so on 1% grade the 1500 ton train is dropping at rate of 0.6 ft/sec

which is 1500tons X 2000 lbs/ton X 0,6 ft/sec = 1.8e6 ft-lbs/sec ,
divide by 550 ft-lbs.sec / hp = 3272 hp

rolling resistance of rail wheels should be in range of 0.002X load,
http://en.wikipedia.org/wiki/Rolling_resistance
3 tons for this train
so we'll subtract that work
3 tons X 2000lbs/ton X 60 ft/sec = 3.6e5 ft-lbs/sec ,
divide by 550 = 654 hp
subtract that from 3272

leaving 2618 hp, or 1953 kilowatts available .

Interesting.

Remember that railcar couplers are not rigid but have a little bit of travel called "slack action".
Back in the dark ages when i rode locomotives the engineers preferred to keep the train stretched out so as to avoid that "clunk" when the couplers moved,
If you let the train compress when starting downhill, the last car gets a "whack" because the front of the train is already moving slower than the rear when the slack gets all used up, so the last car must decelerate quickly. It's like cracking a whip. You'll hear it in slow moving trains in switchyards or at crossings, a loud crash that progresses down the row of cars when engineer reverses direction .

So, in order to give his passengers a comfortable ride the engineer might waste a bit of that kinetic energy in order to keep the couplers in tension .
He'd apply the train brakes which are separate from the locomotive brakes,

Train Slack When running a long train, slack enters into the equation when talking about train control. The old saying goes, "either you control slack or it controls you."When two cars are coupled, there is not a rigid connection there. There is space within the coupler which closes when you pull on the car, and opens when the train in front of the car is going slower than the car is. Many modern cars also have cushion couplers (of different names). When the car is coupled, or slack is run in, the coupler has a shock absorber to avoid damage to the freight, or discomfort to the passengers. These couplers add to the amount of slack in the train...On short trains, like your typical 5-10 car local freight, slack isn't much of a concern. On passenger trains, and long freights, it is something you have to always consider.

A hard "run out" of slack will break your train. The coupler or draft gear will break, and your run will be over until you fix it. This is not a good time to find out you're on bad terms with your co-workers... A hard "run-in" of slack, at best, will make you feel like the locomotive was rear ended by a train, and at worst will cause cars to derail. ...

.....
Using Locomotive or Dynamic Brakes: When train make-up and road characteristics require slack to be bunched, reduce power gradually to allow slack to drift in. Continue to bunch slack by gradual use of dynamic or independent brake.

A quick or heavy application or the independent or dynamic brake will reduce the speed of the [front of the - jh] train more rapidly than the rear. This will result in a slack run-in which can be quite severe. This can cause a rapid heavy buff loading on the rear units of the consist and could cause them to jackknife.

To avoid sliding wheels, dynamic and independent brakes can't be used at the same time. When train brake and dynamic brake are used at the same time, press / to release the locomotive brake. The standard range dynamic brake develops its maximum retarding force between 6 and 23 MPH. Dynamic brakes must always be applied and released gradually.
...

http://www.ovsrails.com/OVSTI/AdvancedTrainManual.html

 

[rollingstein]

in order to give his passengers a comfortable ride the engineer might waste a bit of that kinetic energy in order to keep the couplers in tension .
He'd apply the train brakes which are separate from the locomotive brakes,

Interesting. I always thought that on downhill gradients engineers relied more on the locomotive's dynamic braking than use the air brakes. Because of the dangers of overheated wheels & gradual brake fade. Also because once you did a brake pipe reduction the reservoirs on each car cannot replenish until the brakes were released (I hope I'm getting the details right)

i.e. With train brakes to get increasing braking force on a long downhill run the engineer would have to do successive brake pipe pressure reductions. Until he reached a point where he cannot do any more reductions & reservoirs on the cars are also empty. i.e. No more brakes.

I thought that for especially freight consists descending the passes without a certain minimum number of operative dynamic brake capable axles in the locomotives was strictly verboten as per the Operating Rules.

 

[jim hardy]

Interesting. I always thought that on downhill gradients engineers relied more on the locomotive's dynamic braking than use the air brakes.

That makes sense for a long downhill run - let the train compress(bunch) then brake it from the front.
In the rolling plains where you're constantly going up and down shorter hills and dales it may be desirable to keep the train stretched out. I remember asking an engineer why he didn't simply coast down the gentle Oklahoma hills, he replied "to keep the train stretched out i apply modest train brakes. Gives passengers a smoother ride". Not long after we heard over the radio of a broken drawbar on a freight a couple hundred miles away.. He grimaced and said "Rough handling. Kids today tsk tsk. "

So my experience is from 1960's and from just watching, not from do-ing. If you work for a railroad you are doubtless more current on operations than i.
But to feel first-hand the inertia of a train is , well, something to experience. It seems of cosmic scale.

old jim

 

[anorlunda]

I thought that for especially freight consists descending the passes without a certain minimum number of operative dynamic brake capable axles in the locomotives was strictly verboten as per the Operating Rules.

Exactly right. A famous case happened in San Bernadino California in 1989. http://en.wikipedia.org/wiki/San_Bernardino_train_disaster
The 69 car freight train on a 3% grade, had insufficient dynamic braking descending from a mountain pass. They tried using air brakes. After the accident, the air brakes and wheels were found melted. That is the result of trying to dissipate too much energy in too small a volume in too little time.

But please, regenerative braking is not the same as dynamic braking. Regenerative suggests sending power back to the grid, or storing it in batteries. Dynamic braking dissipates the energy on board.

It'd be interesting to calculate if a 500 kW hotel load would be enough to dissipate all the regen. braking power generated.

Whoa! 500 kW is far too much. That is 100 saunas. 50 kW for hotel load on a train is already very large. To distribute 500 kW on a train would take unreasonably thick cables or unreasonably high voltages, not to mention fantastic junctions at couplers.

 

[rollingstein]

So my experience is from 1960's and from just watching, not from do-ing. If you work for a railroad you are doubtless more current on operations than i.

Oh, I have no railway experience at all. So your analysis is probably better than mine.

 

[rollingstein]

But please, regenerative braking is not the same as dynamic braking. Regenerative suggests sending power back to the grid, or storing it in batteries. Dynamic braking dissipates the energy on board.

Yes. I understand that. My original post was about regen braking in electric locos. Not dynamic.

Even here if you have a diesel loco & try feeding your excess power during braking into hotel load rather than burn it away in a resistor grid that is regenerative braking IMO and not dynamic.

 

[rollingstein]

Whoa! 500 kW is far too much. That is 100 saunas. 50 kW for hotel load on a train is already very large. To distribute 500 kW on a train would take unreasonably thick cables or unreasonably high voltages, not to mention fantastic junctions at couplers.

See page 23 of the RFQ I posted above. It clearly mentions 500 kVA as the peak load of the hotel load module. Granted that peak might be overkill but yet it seems they would size cables for that load? What gives?

I quote:

"The DC link voltage shall be used as input to the hotel load inverter. Major operating parameters of the hotel load module are listed as below...Maximum rated output power: 500 KVA at 0.8 – 1.0 inductive P.F...The control system shall have such a provision that, at any time, not-in-use hotel load power (residual from allocated 500 KVA) shall be used for traction..."

 

[anorlunda]

I stand corrected. It still sounds like a crazy amount.

If they distribute it at 130 v, 500 KVA is 3846 amps. At 600 v it is 833 amps. Higher voltages would be unsafe. If a coupler separated while carrying that much current, it would cause major arcing.

OOOO gauge wire (0.46 inch diameter, and the biggest gauge in the table) has a max rated load current of 302 amps.

Stated another way, 500 KVA is roughly the size of the load of the MGM Grand Hotel in Vegas, or the hotel load of the biggest cruise ships with 6000 passengers. But we are talking about a train with cars that can be easily decoupled.

You cited your source correctly, but I suspect that the author of that RFP misplaced a decimal point. If not, I have egg on my face.

 

[rollingstein]

You cited your source correctly, but I suspect that the author of that RFP misplaced a decimal point. If not, I have egg on my face.

Here's a different way to look at the situation: The load assessment per coach. Your estimate could be low based on this paper I dug up which says:

"Rajdhani, Shatabdi, Garibrath are fully air conditioned trains and require more power. Electrical load per coach for EOG system is in the range of 40 KVA for passenger coaches, 60 KVA for pantry car,.... Average AC coach load in summer is 42KW..."

Ergo, with say 18 coaches 500 kVA will be easily exceeded. Do these loads sound reasonable?

 

[rollingstein]

If they distribute it at 130 v, 500 KVA is 3846 amps. At 600 v it is 833 amps. Higher voltages would be unsafe. If a coupler separated while carrying that much current, it would cause major arcing.

Your point about amperage is very valid. I have seen a multi strand industrial feeder cable at 440 V thick as my wrist & it was rated for 250 A. And I've seen a cable fault and that was a scary sight.

It'd mighty brave to string that sort of cable between coaches of a train.

We need to dig deeper into this...