Medium voltage DC for affordable electrification

[squizzler]

According to International Rail Journal France and Netherlands are considering what to do with their 1500vdc networks. Rather than going for the obvious and upgrade to 25kv ac they are looking at intermediate voltages that reduce the current and size of conductors.

These voltages strike me as eminently suitable for the problematic electrification of Britain's constrained network. The nature of DC vs AC has a bearing on the clearances needed for a given voltage and favours DC. 25kv is only the average (RMS) of the alternating current, the peak is nearly 40kv. 9kv means only a quarter of the potential and should allow considerable clearance reductions. This technology seems to me a shoe-in for third rail network conversion.

Considering we got our 25kv system from France it seems worthwhile to keep an eye on what SNCF are playing with now. If we had to start electrification again right now I doubt we would be using alternating current.

Further options explored include the application of superconductors to power distribution, trackside power storage to clip peak power demands on the grid, along with the inevitable stuff on battery trains and hydrogen traction.

Laying the foundations for energy-efficient traction

While rail is generally considered an energy-efficient form of transport, there is still significant potential for the industry to reduce power consumption and particulate emissions. Keith Barrow looks at some of the technologies that could help infrastructure managers and operators to deliver cost and energy savings.

While the trend in Europe has long been towards standardisation on ac electrification, upgrading of legacy dc systems is now seen as an alternative way forward in some countries. In the Netherlands, where switching the 1.5kV dc system to 25kV ac was given serious consideration in the 1990s, the Ministry of Infrastructure and infrastructure manager ProRail are currently considering converting the network to 3kV dc.

SNCF Network says power electronics technology has reached a state of maturity where 9kV dc is a viable option for rail electrification. High-voltage dc power converters, solid state dc circuit breakers, the availability of medium-voltage drives for industrial motors and the development of Silicon Carbide semi-conductors all support a shift to MVDC.

Superconducting cables are another traction innovation under evaluation in France. Superconducting materials exhibit zero resistance when cooled below a critical temperature, offering potentially significant energy savings. The first trials with superconducting cables on a passenger railway were carried out on Japan’s 1.5kV dc Sunzu Line in 2015.

Trackside energy storage systems (ESSs) are another option for improving the energy efficiency of operations on electrified lines, although such solutions face a number of challenges in the railway environment.

 

[hexagon789]

According to International Rail Journal France and Netherlands are considering what to do with their 1500vdc networks. Rather than going for the obvious and upgrade to 25kv ac they are looking at intermediate voltages that reduce the current and size of conductors.

These voltages strike me as eminently suitable for the problematic electrification of Britain's constrained network. The nature of DC vs AC has a bearing on the clearances needed for a given voltage and favours DC. 25kv is only the average (RMS) of the alternating current, the peak is nearly 40kv. 9kv means only a quarter of the potential and should allow considerable clearance reductions. This technology seems to me a shoe-in for third rail network conversion.

Considering we got our 25kv system from France it seems worthwhile to keep an eye on what SNCF are playing with now. If we had to start electrification again right now I doubt we would be using alternating current.

Further options explored include the application of superconductors to power distribution, trackside power storage to clip peak power demands on the grid, along with the inevitable stuff on battery trains and hydrogen traction.

Laying the foundations for energy-efficient traction

I'd always understood that a.c. electrification and a.c. treaction motors were more efficient than d.c.

 

[swt_passenger]

I'd always understood that a.c. electrification and a.c. treaction motors were more efficient than d.c.

You have to consider separately the external power supply transmission system and the internal motor control. You can still have all the benefits of AC traction motors with power from DC overhead.

 

[hexagon789]

You have to consider separately the external power supply transmission system and the internal motor control. You can still have all the benefits of AC traction motors with power from DC overhead.

I don't doubt it, but isn't a.c. electrification more efficient over longer distances and wouldn't using a.c. electrification with a.c. traction motors be more efficient and cheaper than going from d.c. electrification to a.c. motors?

 

[pdeaves]

You can still have all the benefits of AC traction motors with power from DC overhead.

I think most new third rail (DC) stock has AC motors. You 'just' need an inverter on board.

 

[pdeaves]

I don't doubt it, but isn't a.c. electrification more efficient over longer distances and wouldn't using a.c. electrification with a.c. traction motors be more efficient and cheaper than going from d.c. electrification to a.c. motors?

AC motors tend to work with varying frequency, so an AC input doesn't help unless it is converted to DC first, then it can be converted to whatever you want.

 

[swt_passenger]

I don't doubt it, but isn't a.c. electrification more efficient over longer distances and wouldn't using a.c. electrification with a.c. traction motors be more efficient and cheaper than going from d.c. electrification to a.c. motors?

Transmission outside the train is more efficient with higher AC voltage, but within the train itself it could actually be more efficient supplied with suitable DC, because there’s already a DC stage between the single phase AC incoming supply and the inverter derived 3 phase AC Traction package. You lose a transformer/rectifier if the OHLE is a usable DC voltage.

 

[hexagon789]

AC motors tend to work with varying frequency, so an AC input doesn't help unless it is converted to DC first, then it can be converted to whatever you want.

So it needs to be converted to DC first then.

Transmission outside the train is more efficient with higher AC voltage, but within the train itself it could actually be more efficient, because there’s already a DC stage between the single phase AC incoming supply and the inverter derived 3 phase AC Traction package. You lose a transformer/rectifier if the OHLE is a usable DC voltage.

Well that's explains something then as I thought it went AC to AC rather than AC-DC-AC.

 

[swt_passenger]

Well that's explains something then as I thought it went AC to AC rather than AC-DC-AC.

It all has to work in both directions as well if regenerative braking is a feature. So where people refer to rectifiers and inverters they have to work in both directions, so the devices are in reality dual purpose devices referred to as “converters”. When powering the normal flow is AC in through a transformer to some lower voltage AC, then rectified to DC, then inverted to 3 phase AC. Braking produces 3 phase AC which is rectified to DC, then inverted to single phase AC then transformed up to the OHLE voltage.

I think one of the reasons this isn’t so obvious is if people discuss the additional equipment that needs fitting to a DC supplied EMU to make it AC capable (eg making 707s into go anywhere units), they nearly always mention the pantograph and transformer, but almost never mention the aforementioned converter...

 

[edwin_m]

Didn't BR establish that 6.25kV AC didn't need extra clearance compared with 1500V DC? After more experience they decided that converting the 6.25kV to 25kV was less hassle than continuing with a dual-voltage system, which probably means that most of the original clearances were OK for 25kV too?

If this is so then it suggests the electrical clerances are a red herring and the problem is more about physically accommodating the wire regardless of voltage - and introducing yet another non-standard feature to the rail network. However this may be complicated by the much larger and probably unnecessary structural clearances dictated by recent standards changes.

Transmission outside the train is more efficient with higher AC voltage, but within the train itself it could actually be more efficient supplied with suitable DC, because there’s already a DC stage between the single phase AC incoming supply and the inverter derived 3 phase AC Traction package. You lose a transformer/rectifier if the OHLE is a usable DC voltage.

However I suspect you still need some pretty chunky filtering as the DC link will include lots of variable frequency AC components created by the converters, and these have to be reduced to a level that won't affect the signalling.

 

[swt_passenger]

Didn't BR establish that 6.25kV AC didn't need extra clearance compared with 1500V DC? After more experience they decided that converting the 6.25kV to 25kV was less hassle than continuing with a dual-voltage system, which probably means that most of the original clearances were OK for 25kV too?

If this is so then it suggests the electrical clerances are a red herring and the problem is more about physically accommodating the wire regardless of voltage - and introducing yet another non-standard feature to the rail network. However this may be complicated by the much larger and probably unnecessary structural clearances dictated by recent standards changes.

However I suspect you still need some pretty chunky filtering as the DC link will include lots of variable frequency AC components created by the converters, and these have to be reduced to a level that won't affect the signalling.

Yes all that makes sense. I don’t actually see it happening, my main aim was only to clarify how power transmission and use on the train are separable matters.

 

[hexagon789]

It all has to work in both directions as well if regenerative braking is a feature. So where people refer to rectifiers and inverters they have to work in both directions, so the devices are in reality dual purpose devices referred to as “converters”. When powering the normal flow is AC in through a transformer to some lower voltage AC, then rectified to DC, then inverted to 3 phase AC. Braking produces 3 phase AC which is rectified to DC, then inverted to single phase AC then transformed up to the OHLE voltage.

I think one of the reasons this isn’t so obvious is if people discuss the additional equipment that needs fitting to a DC supplied EMU to make it AC capable (eg making 707s into go anywhere units), they nearly always mention the pantograph and transformer, but almost never mention the aforementioned converter...

I really do appreciate when people explain things like this, so thanks.

Yes, I agree - the way people often referred to this needing this and that needing that often hampers actually understanding and appreciating what something really involves.

 

[AM9]

Didn't BR establish that 6.25kV AC didn't need extra clearance compared with 1500V DC? After more experience they decided that converting the 6.25kV to 25kV was less hassle than continuing with a dual-voltage system, which probably means that most of the original clearances were OK for 25kV too?

If this is so then it suggests the electrical clerances are a red herring and the problem is more about physically accommodating the wire regardless of voltage - and introducing yet another non-standard feature to the rail network. However this may be complicated by the much larger and probably unnecessary structural clearances dictated by recent standards changes. ...

The application of ac in areas where there were frequent bridges and tunnels with minimum clearance was because the mandated clearance for 25kV ac was about 11 inches. After problems with some of the on-train transformers and associated switchgear, the minimum clearance for 25kV was re-evaluated and reduced in steps to ISTR about 5 inches. This enabled the troublesome 6.25Kv OLE sections to be uprated to 25Kv. Much of the 6.25Kv was on the GEML where the original 1500VDC insulators were used for 6.25 kV, - these could not be used on 25kV so were changed.
The clearance issues are now involving passenger safety especially in stations where the horns of pantographs may be too close.

 

[big all]

on the southern there was a substation at every 2miles approximately and a tp hut [track parallel hut]in between to spread the load around like a ring main
they then around 1990 in redhill area at least doubled the substations to around every mile in prep for channel tunnel running
so whilst clearance is far less for dc the volt loss is great so not the only consideration

 

[apk55]

Unless you are talking about a totally isolated line with no possibilities of connections to the rest of the network and sufficient demand to justify a reasonable pool of stock then you are best sticking with 25KV AC. Even with 1500V DC you require sub stations every 3 to 6 miles (it is the return voltage drop via the rails that determines that distance - too much and you run the risk of electrocution and corrosion problems). Otherwise you will end up with interface problems where two systems meet requiring the use of dual voltage stock or an isolated part of the station

 

[edwin_m]

The clearance issues are now involving passenger safety especially in stations where the horns of pantographs may be too close.

There are two rather separate issues as I understand it.

The clearance between platforms and wires is a passenger safety issue but the recommended clearances can be reduced subject to risk assessment. One measure I wonder about could be a canopy with a deep valance which would help to block any access to pantograph horns and would have the incidental benefit of keeping the rain off the passengers much better than most of the current efforts.

There are also newish standards for clearance between 25kV and structures, but as far as I can see that's not a safety issue as any flashover to an earthed structure is unlikely to endanger anybody. It should be possible to accept lower clearances at the price of a few extra trippings out such as when a pigeon vapourises itself by perching on the wire. This now seems to be possible, for example in Cardiff where NR has insulated the Valley Lines overbridge rather than raising the Valley Lines or lowering the GW main line.

 

[AM9]

There are two rather separate issues as I understand it.

The clearance between platforms and wires is a passenger safety issue but the recommended clearances can be reduced subject to risk assessment. One measure I wonder about could be a canopy with a deep valance which would help to block any access to pantograph horns and would have the incidental benefit of keeping the rain off the passengers much better than most of the current efforts.

There are also newish standards for clearance between 25kV and structures, but as far as I can see that's not a safety issue as any flashover to an earthed structure is unlikely to endanger anybody. It should be possible to accept lower clearances at the price of a few extra trippings out such as when a pigeon vapourises itself by perching on the wire. This now seems to be possible, for example in Cardiff where NR has insulated the Valley Lines overbridge rather than raising the Valley Lines or lowering the GW main line.

In the context of conductors carrying 25kV ac vs. 1500 (or 3000) VDC, the physical spacing is largely determined by safety of movable wire(s) which are subject to deflection from many sources including:

directly from pantograph upward pressure
pantograph oscillation
pantograph transverse movement from sway of the vehicle on which it is mounted
oscillation and travelling waves, especially from multiple EMUs each with raised pantographs
track/wire misalignment caused by ballast creep/catenary support instability
wind effects in exposed areas​

The flashover distance for 25kV or 1500V is a small part of the overall dynamic requirements and at lower voltages there would be little reduction in the amount of overhead infrastructure that needed modification. So lowering the voltage would make little improvement in route clearance, but would have a major power distribution cost impact.

 

[61653 HTAFC]

<snip> This technology seems to me a shoe-in for third rail network conversion.
<snip>

Well played!

 

[edwin_m]

Well played!

Oooh, didn't pick up on that one...

 

[squizzler]

Well played!

Oooh, didn't pick up on that one...

Please stick to the current topic, thank you!

The application of ac in areas where there were frequent bridges and tunnels with
minimum clearance was because the mandated clearance for 25kV ac was about 11 inches. After problems with some of the on-train transformers and associated switchgear, the minimum clearance for 25kV was re-evaluated and reduced in steps to ISTR about 5 inches. This enabled the troublesome 6.25Kv OLE sections to be uprated to 25Kv. Much of the 6.25Kv was on the GEML where the original 1500VDC insulators were used for 6.25 kV, - these could not be used on 25kV so were changed.

Interesting point relating to the reuse of 1.5kV dc insulators for 6.25kV ac. This corresponds closely with the SNCF proposal of using 9kV DC to replace the 1.5kV network. 6.25kV is the Root Mean Square average voltage of AC, the waveform reaches a peak voltage of 8.88kV.

Even with 1500V DC you require sub stations every 3 to 6 miles (it is the return voltage drop via the rails that determines that distance - too much and you run the risk of electrocution and corrosion problems). Otherwise you will end up with interface problems where two systems meet requiring the use of dual voltage stock or an isolated part of the station

The article says operators are looking at upgrading away from 1500V OLE to either 9kV or 3kV dc. Or were you suggesting upgrading the third rail to 1500V (that would almost certainly never be deemed acceptable on safety grounds!)?

Unless you are talking about a totally isolated line with no possibilities of connections to the rest of the network and sufficient demand to justify a reasonable pool of stock then you are best sticking with 25KV AC

I don't think multi-voltage stock is as difficult as it was in the past. At any rate the third rail territory is not so dissimilar from your ideal self contained system as a place to consider higher voltage DC overhead as a replacement to third rail. Also to consider is that the track circuits here will be immunised against DC currents not AC frequencies and going with a higher voltage DC would save the cost of immunisation works.

Of course switching to DC has other advantages besides safety. I understand it becomes cheaper to connect to the grid because the power requirements are shared between all three phases of the mains supply. Whereas 25kv is single phase and has to connect into the uppermost tier of the national grid to avoid unbalancing the phases, a DC supply can connect to a lower tier of the network. I also believe that solid state power electronics for DC voltage conversion are lighter and smaller than the transformers needed for AC, resulting in benefits for rolling stock design.

 

[edwin_m]

I don't think multi-voltage stock is as difficult as it was in the past. At any rate the third rail territory is not so dissimilar from your ideal self contained system as a place to consider higher voltage DC overhead as a replacement to third rail. Also to consider is that the track circuits here will be immunised against DC currents not AC frequencies and going with a higher voltage DC would save the cost of immunisation works.

Of course switching to DC has other advantages besides safety. I understand it becomes cheaper to connect to the grid because the power requirements are shared between all three phases of the mains supply. Whereas 25kv is single phase and has to connect into the uppermost tier of the national grid to avoid unbalancing the phases, a DC supply can connect to a lower tier of the network. I also believe that solid state power electronics for DC voltage conversion are lighter and smaller than the transformers needed for AC, resulting in benefits for rolling stock design.

AC feeders do need a high voltage Grid connection, but they cover a far larger area than DC ones so a suitable place can usually be found without having to lay a lot of extra cables. It would now be possible to install a DC link and inverters in the feeder station too, to balance the load - interconnectors and some other long-distance transmission cables now use DC with similar technology to interface to the AC each end.

It was mentioned somewhere above that the presence of a DC link makes it somewhat easier to power an AC-motored train from a DC supply. As far as I'm aware all modern dual-voltage trains in the UK use a 750V DC link for this very reason, which would mean some extra complication if they were to be fed from a higher-voltage DC supply. I believe the class 323 uses a DC link at 1500V, not surprising as the traction package is based on something developed for the Netherlands.

The transformer and associated equipment (traditionally a rectifier, now some extra conversion electronics) does add weight and take up space on an AC train. This is the main reason why metros generally use DC - the intense service needs a lot of substations but as distances are short the power loss from the lower voltage doesn't matter too much. Using third rail instead of overhead may also allow a smaller tunnel cross-section. These lines are mostly in tunnel or elevated and often with platform screen doors and bottom contact rail like DLR so the safety concerns are much less. Most metros are electrified at 750V but a couple of recent ones have third rail at 1500V. Trams universally use DC at 750V or less for the same reason but also for safety in street running. However for mixed-traffic railways the energy savings and increased performance possible with a 25kV system outweighs the need to carry extra equipment.

 

[HSTEd]

The way to improve third rail installations is to make use of improved power electronics and simply add more substations.

Because we can now use current control and instantaneous inter-tripping thanks to our fancy MOSFETs (made of things like Silicon Carbide and Gallium Nitride) we can have numerous substations that cooperate in powering a train, rather than relying on the two nearest substations.

As the losses in the third rail scale by the square of the current flowing through the rail, by moving the source of some current closer to the train we can achieve drastic reduction in losses whilst remaining within the existing third rail specification.

About the only reason to abandon the third rail spec for overhead DC would be to go to about 12kV, as was proposed in the Soviet Union as a replacement for 3kV and 25kVac in the 80s.
12-25kVdc would be an interesting specification, but as I said, I believe we should instead focus on harnessing the improvements in controlled rectifiers to produce a better third rail spec.

(And yes I truly believe in third rail as many people around here will know! I believe it can be made suitable for the modern era through the use of modern technologies, for example a normally dead segment of third rail in the centre of station platforms that is only lived up if a train is gapped in the platform)

 

[Aictos]

Why not instead of looking at DC, have a look at AC and introduce the same system that Switzerland/Austria/Germany use which is 15kV AC after all it would make sense for cross border services and it is a proven power supply.

Also SNCF already operate locos which operate off this power supply in the form of the TGV POS/Réseau Duplex tri current locos which serve SNCF destinations in Germany and Switzerland plus it improves interoperationality.

Switzerland. Austria, Germany, Netherlands and France all border each other so makes sense to share the same electrification type if 25kV AC is not going to be the proposed electrification type.

 

[squizzler]

AC feeders do need a high voltage Grid connection, but they cover a far larger area than DC ones so a suitable place can usually be found without having to lay a lot of extra cables. It would now be possible to install a DC link and inverters in the feeder station too, to balance the load - interconnectors and some other long-distance transmission cables now use DC with similar technology to interface to the AC each end.

I recall reading something about the ability to use all three phases of grid electricity in the way you describe.But now look at the path the electrical energy must take on an ac system:

3 phase ac --> high voltage dc? --> 25kV ac 1 phase -->(OLE) --> low voltage dc --> low voltage 3 phase ac --> motion!

Whereas on the dc OLE:

3 phase ac --> high voltage dc --> (OLE) --> low voltage dc --> low voltage 3 phase ac --> motion!

It was mentioned somewhere above that the presence of a DC link makes it somewhat easier to power an AC-motored train from a DC supply. As far as I'm aware all modern dual-voltage trains in the UK use a 750V DC link for this very reason, which would mean some extra complication if they were to be fed from a higher-voltage DC supply. I believe the class 323 uses a DC link at 1500V, not surprising as the traction package is based on something developed for the Netherlands.

The transformer and associated equipment (traditionally a rectifier, now some extra conversion electronics) does add weight and take up space on an AC train. This is the main reason why metros generally use DC - the intense service needs a lot of substations but as distances are short the power loss from the lower voltage doesn't matter too much. Using third rail instead of overhead may also allow a smaller tunnel cross-section. These lines are mostly in tunnel or elevated and often with platform screen doors and bottom contact rail like DLR so the safety concerns are much less. Most metros are electrified at 750V but a couple of recent ones have third rail at 1500V. Trams universally use DC at 750V or less for the same reason but also for safety in street running. However for mixed-traffic railways the energy savings and increased performance possible with a 25kV system outweighs the need to carry extra equipment.

The medium voltage dc systems strike me as an excellent compromise between the simplicity and lightness of the low voltage metro systems and the high voltage heavy rail systems. In terms of the size and weight of trains British 'heavy rail' isn't so heavy by global standards. The frequency of trains on our system and the physical constraint of the loading gauge also lie partway between other countries heavy rail and urban rail practices. I would argue that the optimal power system should also occupy a goldilocks voltage such as 12kV dc.

About the only reason to abandon the third rail spec for overhead DC would be to go to about 12kV, as was proposed in the Soviet Union as a replacement for 3kV and 25kVac in the 80s.
12-25kVdc would be an interesting specification, but as I said, I believe we should instead focus on harnessing the improvements in controlled rectifiers to produce a better third rail spec.

I agree 12kV would be even more powerful than the 9kV dc proposed by SNCF which seems to be chosen to be compatible with the existing 1500V OLE.

I disagree however that 3rd rail is an appropriate system for further development and expansion. The originally linked article advocates higher voltages on existing systems which I think would be beyond the pale for the expansive British Southeastern system.

 

[edwin_m]

I recall reading something about the ability to use all three phases of grid electricity in the way you describe.But now look at the path the electrical energy must take on an ac system:

3 phase ac --> high voltage dc? --> 25kV ac 1 phase -->(OLE) --> low voltage dc --> low voltage 3 phase ac --> motion!

Whereas on the dc OLE:

3 phase ac --> high voltage dc --> (OLE) --> low voltage dc --> low voltage 3 phase ac --> motion!

I've already pointed out that the inverters at feeder stations don't appear to be necessary and as far as I know not even new ones have them. It is simply something that could be done to address phase imabalance if it was ever needed.

So the more usual AC sequence is:
3 phase ac --> 25kV ac 1 phase -->(OLE) --> low voltage dc --> low voltage 3 phase ac --> motion!

Converting high (actually medium) voltage DC to low voltage would need another set of electronics. Your DC sequence should read:
3 phase ac --> medium voltage dc --> (OLE) --> low voltage dc --> low voltage 3 phase ac --> motion!

Now both sequences have the same number of arrows. And remember, other things being equal, resistance losses vary with the square of the voltage and the resulting voltage drop also reduces the scope for regeneration. Every electric transmission system in the world says that the reduced losses from using a higher voltage outweigh the losses in the transformers from stepping it up and down.

 

[ABB125]

One thing that seems rather wasteful to me is converting the AC overhead supply to DC, and then AC again to power the motors. Is it not possible to simply transform the 25kV to lower voltage and send it directly to the motors?

 

[squizzler]

One thing that seems rather wasteful to me is converting the AC overhead supply to DC, and then AC again to power the motors. Is it not possible to simply transform the 25kV to lower voltage and send it directly to the motors?

That's what happened in the good old days with dc motors. When talking about ac drive on trains it means there are electronics that produce - normally across three wires (phases) leading to the motors - pulses of power (both positive and negative, hence alternating current) that drive the magnets within the motor. These pulses are of variable voltage and frequency and tailored to what the drive (or braking) is needed at the time.

ac from the national grid is also three phase but is of fixed voltage and at precisely 50 hz frequency so no good for driving a train which has to run at variable speeds.

 

[Greybeard33]

AC feeders do need a high voltage Grid connection, but they cover a far larger area than DC ones so a suitable place can usually be found without having to lay a lot of extra cables.

Really? What about:

• The costly ground level extension lead from Heyrod grid feed point to Manchester Victoria?
• Likewise from Thingley Junction to east of Chippenham?
• Reinstatement of Kettering to Market Harborough electrification, primarily to avoid yet another extension lead from Braybrooke grid feed point?

The volts drop at 9kV DC is only one-twelfth of that at 750V DC for the same power, so substations can be 12 times further apart (16 times for 12kV).

The need for a direct grid connection can make isolated "islands" of 25kV electrification, e.g. extensions/infill of third rail electrification, prohibitively costly. Whereas DC OLE substations, being a balanced load on the three phase supply, can be fed at lower voltage from the local DNO network, at much lower cost.

It would now be possible to install a DC link and inverters in the feeder station too, to balance the load - interconnectors and some other long-distance transmission cables now use DC with similar technology to interface to the AC each end.

A DC link and inverters cannot be used to convert a single phase load into a balanced three phase load. This is because single phase power flow is inherently pulsating, falling to zero at each zero crossing of the waveform, whereas balanced three phase power flow is constant and continuous. What is needed is a Static VAR Compensator, which is a big, costly installation because it uses large capacitors to temporarily store energy and so smooth out the power flow. Power electronics devices cannot replace the energy storage function.

In contrast, a DC transmission link between isolated three phase networks carries a continuous power flow and so does not need energy storage devices.

It was mentioned somewhere above that the presence of a DC link makes it somewhat easier to power an AC-motored train from a DC supply. As far as I'm aware all modern dual-voltage trains in the UK use a 750V DC link for this very reason, which would mean some extra complication if they were to be fed from a higher-voltage DC supply.

A train can carry a DC-DC converter to enable different DC supply voltages to be used, e.g. 9kV OLE and 750V third rail. The converter can operate at a higher frequency than 50Hz, enabling the power transformer to be smaller and lighter than for an AC train.

DC OLE does not need the complexities of neutral sections to separate areas fed from different supply phases.

25kV 50Hz single phase was a good compromise with mid 20th century technology, but the advances in power electronics since have eroded its advantages versus high voltage DC OLE.

 

[edwin_m]

Really? What about:

• The costly ground level extension lead from Heyrod grid feed point to Manchester Victoria?
• Likewise from Thingley Junction to east of Chippenham?
• Reinstatement of Kettering to Market Harborough electrification, primarily to avoid yet another extension lead from Braybrooke grid feed point?

The first and third of these are partly the consequence of there being a long lead time on grid connections but primarily because of bad planning leading to cancellation of the electrification that was going to run past the feeder. I believe Thingley was always going to have an extension lead to Wootton Bassett so it could feed towards Swindon and Parkway while the OLE leading to Thingley itself was isolated for maintenance etc.

The need for a direct grid connection can make isolated "islands" of 25kV electrification, e.g. extensions/infill of third rail electrification, prohibitively costly. Whereas DC OLE substations, being a balanced load on the three phase supply, can be fed at lower voltage from the local DNO network, at much lower cost.

I've already pointed out this could be done with 25kV too if there was a need to, using inverters.

A DC link and inverters cannot be used to convert a single phase load into a balanced three phase load. This is because single phase power flow is inherently pulsating, falling to zero at each zero crossing of the waveform, whereas balanced three phase power flow is constant and continuous. What is needed is a Static VAR Compensator, which is a big, costly installation because it uses large capacitors to temporarily store energy and so smooth out the power flow. Power electronics devices cannot replace the energy storage function.

So how does an modern traction package generate three-phase AC from a DC link on board a train? In any case an inverter feeder station would be fed with three-phase just like a DC substation.

A train can carry a DC-DC converter to enable different DC supply voltages to be used, e.g. 9kV OLE and 750V third rail. The converter can operate at a higher frequency than 50Hz, enabling the power transformer to be smaller and lighter than for an AC train.

DC OLE does not need the complexities of neutral sections to separate areas fed from different supply phases.

25kV 50Hz single phase was a good compromise with mid 20th century technology, but the advances in power electronics since have eroded its advantages versus high voltage DC OLE.

That may be true but you are introducing yet more complexity and moving away from compatibility with the historic system which is now standard internationally. Isn't the railway hobbled enough already by non-standard features which restrict the ability to use any rolling stock anywhere?

 

[Greybeard33]

The first and third of these are partly the consequence of there being a long lead time on grid connections but primarily because of bad planning leading to cancellation of the electrification that was going to run past the feeder. I believe Thingley was always going to have an extension lead to Wootton Bassett so it could feed towards Swindon and Parkway while the OLE leading to Thingley itself was isolated for maintenance etc.

So-called independent feeders, such as that from Thingley to Wooton Bassett, are another costly consequence of reliance on a small number of widely spaced connections to the Supergrid. These feeders are needed to enable sections of the OLE to be bypassed when they are isolated for maintenance. With DC, it might have been feasible to provide a DNO feed point near Wooton Bassett, to obviate the need for the independent feeder.

The wide spacing between Supergrid supply points also necessitates the use of autotransformer stations, with long ATF feeder cables from the grid supply point adding to costs.

It would be prohibitively costly to provide dedicated tappings off the 400kV Supergrid especially for railway use, which is why some of the supply points have to be in relatively inconvenient locations like Heyrod and Melksham.

I've already pointed out this could be done with 25kV too if there was a need to, using inverters.

And I already pointed out this would need a SVC, not an inverter. Without capacitor storage to smooth out the pulsating single phase load, a 3-phase rectifier> DC link> 50 Hz single phase inverter arrangement would impose an unbalanced load on the 3-phase supply, drawing power from only one or two of the supply phases. No advantage over connecting a single phase transformer across two phases, as is actually done at grid supply points.

So how does an modern traction package generate three-phase AC from a DC link on board a train? In any case an inverter feeder station would be fed with three-phase just like a DC substation.

On board the train, full wave rectification of the 50Hz single phase supply inevitably creates a 100Hz AC ripple superimposed on the DC link. I do not know the detail of how traction packages cope with this, but in principle there are two approaches:

• Capacitive/inductive energy storage to smooth the ripple on the DC link, so that the traction package can maintain a steady voltage to the motors
• Allow the AC motor voltage to be modulated at 100Hz (not good for the motor in respect of mechanical vibration/bearing wear, but the inductance of the motor windings will help to smooth the current).

Or a combination of the two. I understand AC traction motors operate at a higher frequency than 100Hz, so even with the second approach the traction package can generate a 3-phase output from the fluctuating DC link.

That may be true but you are introducing yet more complexity and moving away from compatibility with the historic system which is now standard internationally. Isn't the railway hobbled enough already by non-standard features which restrict the ability to use any rolling stock anywhere?

Only more complex if the rolling stock needs to be triple voltage capable. And I am afraid that the "go anywhere" bird has long since flown the nest! But I concur that standardisation (and accumulated knowledge/experience) are strong reasons to stick with "the devil(s) you know",
unless it can be demonstrated that there are overwhelming benefits from introduction of yet another electrification standard.