What’s more efficient: overhead or 3rd rail electrification?

[Taunton]

This one has rumbled on here for a long time, including by some of the contributors above! One of the issues is proponents of either system extolling all the virtues they can come up with, while staying silent about any downsides.

3rd rail is a sight easier, quicker and cheaper to install. Having watched it actually go in one weekend I'll offer by a factor of 10 times. It doesn't suit high speed running (unless you are doing Waterloo to Bournemouth it seems), but does suit Metro systems, which worldwide seem to use little else, even new ones - why does the Dubai Metro, for example, brand new and a substantial system, not use OHLE?

End of the day it's horses for courses, which is why saying Reading to Gatwick, with three existing third rail sections interspersed with two non-electrified bits, should have the latter done with OHLE, just seems silly. About as silly as a little 3rd rail bit over the Border Bridge!

 

[WirralLine]

Composite conductor rail, mostly aluminium for lower resistance with a steel layer on top to reduce wear from the shoes. This presumably allowed fewer substations while keeping voltage drop within limits. I don't think it was used anywhere else so I assume there were problems with it.

Pretty sure there's a lot of this rail in use on the Merseyrail network. It's much thinner, and is a light silver/grey colour on the bottom 3/4 of it. The majority of the Hooton-Chester/Ellesmere Port extension of 3rd rail in the 90s uses it, which I believe was done on the cheap although it can be found in places all over the network.

The Chester end has suffered from low power since day 1 and has had "E" restriction boards up for god knows how many years. Pretty sure it was being discussed in another thread recently that the whole south section of the Wirral Lines are fed from Bromborough grid, with a substation at Hooton and Capenhurst only.

 

[edwin_m]

3rd rail ... doesn't suit high speed running (unless you are doing Waterloo to Bournemouth it seems), but does suit Metro systems, which worldwide seem to use little else, even new ones - why does the Dubai Metro, for example, brand new and a substantial system, not use OHLE?

100mph is the maximum on the Bournemouth line, which isn't really high speed these days.

As mentioned above, third rail for a metro can significantly reduce the cross-section of the tunnel, which has a big impact on costs, and a high frequency of trains making numerous stops suits a system where the transformer is in the substation not on the train. Another factor is that metros don't generally have level crossings and maintenance usually involves isolating the power during non-traffic periods, so there is less electrical hazard than on a third rail main line railway.

 

[JamesT]

This one has rumbled on here for a long time, including by some of the contributors above! One of the issues is proponents of either system extolling all the virtues they can come up with, while staying silent about any downsides.

3rd rail is a sight easier, quicker and cheaper to install. Having watched it actually go in one weekend I'll offer by a factor of 10 times. It doesn't suit high speed running (unless you are doing Waterloo to Bournemouth it seems), but does suit Metro systems, which worldwide seem to use little else, even new ones - why does the Dubai Metro, for example, brand new and a substantial system, not use OHLE?

End of the day it's horses for courses, which is why saying Reading to Gatwick, with three existing third rail sections interspersed with two non-electrified bits, should have the latter done with OHLE, just seems silly. About as silly as a little 3rd rail bit over the Border Bridge!

The Barcelona Metro, which has lines varying in age from 10 years old to 100 years old is all overhead conductors at 1200-1500V DC.
The Delhi Metro which the oldest bit is only 25 years old chose 25kV AC overhead, so I don't think you can generalise anything from whether a particular system chose AC or DC, it most likely depended on what the contractor building it was most familiar with.

 

[QueensCurve]

Thanks both for the technical stuff. Weren't the contact wires used on the Woodhead DC Overheads considerably thicker and heavier than what we use on the 25kv 50hz, too?

Yes. Last time I looked, many years ago, the contact wire was noticeably thicker on platforms 1-4 at Manchester Piccadilly than the other platforms.

 

[Zomboid]

The Barcelona Metro, which has lines varying in age from 10 years old to 100 years old is all overhead conductors at 1200-1500V DC.
The Delhi Metro which the oldest bit is only 25 years old chose 25kV AC overhead, so I don't think you can generalise anything from whether a particular system chose AC or DC, it most likely depended on what the contractor building it was most familiar with.

I think you probably can generalise, but there are exceptions.
Crossrail being one, though it's clear in that case that making the tunnels big enough to work with 25kV would be simpler than dual voltage trains switching over at either end - I bet it was considered at option selection though.

 

[Taunton]

Yes. Last time I looked, many years ago, the contact wire was noticeably thicker on platforms 1-4 at Manchester Piccadilly than the other platforms.

Yes, that's because for original 1,500V DC the current is much higher than 25kV. I believe it was said that when the wiring was replaced on the GEML a significant proportion of the project cost was covered from the scrap value of the old onetime DC wiring compared to the replacing AC wiring.

 

[edwin_m]

I think you probably can generalise, but there are exceptions.
Crossrail being one, though it's clear in that case that making the tunnels big enough to work with 25kV would be simpler than dual voltage trains switching over at either end - I bet it was considered at option selection though.

Tunnels bored by TBM have to be circular, and depending on the train structure gauge and evacuation walkway dimensions there may be room for OLE without making it any bigger. Also, for higher speed routes, extra free cross-section reduces pressure effects and must save some amount of power consumption. So for these types of metro an AC overhead wire may not add to the cost.

A cut and cover tunnel may be a rectangular section, in which case overhead wire may have more of a cost impact by forcing the track to be a bit lower.

 

[Energy]

Metros are different to the main line. DC allows the tunnels to be built smaller, and more frequent distribution sites can be advantageous in highly intensive service areas.

I wouldn't say it's advantageous, but the frequent feeder stations aren't as much of a concern for metros, where the distance covered is less. AC electrification is more efficient over long distances, but requires heavy transformers and rectifiers on board.

Didn't a NIMBY suggest 3rd Rail for Sydney Gardens and around Bath Spa?

Possibly, though an electrification design was agreed upon in 2015. [source]

Image Description - A 3D model of the cantilever design for electrification through Bath.

 

[Sun Chariot]

Yes. Last time I looked, many years ago, the contact wire was noticeably thicker on platforms 1-4 at Manchester Piccadilly than the other platforms.

I now need to peruse my 35mm photos taken at Manchester Piccadilly from 1989-1990! I stood there often enough but I completely failed to spot the subtle difference. Face palm!

 

[GraphitProjekt]

The Third Rail south of the river was basically the LSWR system, the LBSCR used a DC Overhead which went as far south as Coulsdon North, as well as a good number of other routes,

As others have said, it was a 6.6kV AC system at 25Hz. It seems from around 1900 there were two options for electrification: low voltage DC (between ~600 and 3000V) on conductor rails or OLE, or high(er) voltage AC (between 6kV and 15kV) but using bespoke power stations and grid infrastructure to deliver lower frequency (than mains) AC. I can't claim to understand why, but it seem that even back in the day you could get a decent traction motor out of single phase AC specifically at a low frequency being fed directly to a "universal" motor (whatever that it...). In Britain the low frequency high voltage AC method only was done by LBSCR and also the short line between Lancaster and Morcambe/Heysham. However it became the national standard in Germany/Switzerland/Austria and Sweden/Norway using 15kV 16.7Hz (or 16 and two thirds for Scandinavia). It's also used between Washington DC and New York City as well as the suburban lines around Philadelphia at 12kV 25Hz.

However, I can't think of any totally new main line electrification scheme

No doubt somebody can find an exception but I don't believe there are any major ones.

I guess it's not really mainline but there's a new regional suburban system being built in Montreal at the moment (Réseau express métropolitain or REM) that is 1.5kV DC OLE. It's mostly all brand new alignments but does involve taking over and converting a previous suburban line that was 25kV AC (the Deux-Montagnes line, the only electric heavy rail route in Canada)

Pretty sure there's a lot of this rail in use on the Merseyrail network. It's much thinner, and is a light silver/grey colour on the bottom 3/4 of it. The majority of the Hooton-Chester/Ellesmere Port extension of 3rd rail in the 90s uses it, which I believe was done on the cheap although it can be found in places all over the network.

The Chester end has suffered from low power since day 1 and has had "E" restriction boards up for god knows how many years. Pretty sure it was being discussed in another thread recently that the whole south section of the Wirral Lines are fed from Bromborough grid, with a substation at Hooton and Capenhurst only.

To my knowledge, the electrification south of Rock Ferry introduced one new 33/11 grid connection at Bromborough as well as DC substations at Bromborough, Hooton, Ellesmere Hurst, Mollington and a paralleling hut at Port Sunlight. From 2019 a power supply upgrade was commenced for the whole Merseyrail system which introduced a new traction substation at Capenhurst, converted Port Sunlight from TPH to a full substation and added a second 33/11kV connection at Bromborough.

Metros are different to the main line. DC allows the tunnels to be built smaller, and more frequent distribution sites can be advantageous in highly intensive service areas.

It's probably possible that the number of transformers and rectifiers to be purchased is fewer if you can put them in the substations rather than on each train for a railway with short distances and very high frequencies (which is another way of saying "metro"). I have also heard it said that the higher contact area on the conductor you get for 3rd rail is mildly advantageous for lots of trains accelerating all at once. Probably not a very big difference though.

This also makes circuit protection more challenging, as the resistance may limit the current drawn by a short a long way from a substation to something similar to what is seen normally if a few trains happen to be accelerating at once. The fault current in a 25kV system will be much more than the normal current.

I have heard it said that this is actually a significant limiting factor in substation distancing on the 750V system. So much so that Bournemouth to Weymouth implemented some kind of complex active control system whereby a breaker opening at one end of the feeding area automatically communicates with the one at the other end to make sure they both open. And this gave them some marginal gains in substation spacing

 

[zwk500]

I think you probably can generalise, but there are exceptions.
Crossrail being one, though it's clear in that case that making the tunnels big enough to work with 25kV would be simpler than dual voltage trains switching over at either end - I bet it was considered at option selection though.

3rd Rail was suggested by an MP at committee stage for the Maidenhead bridge, although that was for visual impact reasons not anything operational.

 

[Sorcerer]

However it became the national standard in Germany/Switzerland/Austria and Sweden/Norway using 15kV 16.7Hz (or 16 and two thirds for Scandinavia).

What's interesting to note about the 15kV 16.7Hz system is that aside from needing a larger step-down transformer on-board the train and relying on electromechanics rather than power electronics, it still pretty much shares all the same advantages as 25kV 50/60Hz over DC systems which is why it was best suited for electrification in the aforementioned countries and why new infrastructure such as the new base tunnels still use it.

 

[edwin_m]

I can't claim to understand why, but it seem that even back in the day you could get a decent traction motor out of single phase AC specifically at a low frequency being fed directly to a "universal" motor (whatever that it...).

A universal motor is basically a DC motor, feeding current into stator coils and via a commutator into armature coils so that the magnetic field switches direction to force the armature to move. The coils form an inductor, so will have an impedance that increases at higher AC frequencies. This means that at higher frequencies more of the power is wasted as heat in the coils, not only reducing efficiency but also limiting the motor power according to how well it can be cooled. I guess this was thought to be intolerable at 50Hz or 60Hz grid frequencies so a lower frequency was used (although at that time grid frequencies probably varied quite a bit place to place).

Grid frequency electrification wasn't an option until the advent of rectifiers suitable for use on board, which would convert the AC into DC to feed it into a DC motor. These were initially mercury arcs but rapidly superseded by semiconductors.

Finally, the rotation speed of an AC motor is closely tied to the supply frequency. This obviously isn't ideal for a train that has to run at a range of speeds, so they didn't become widespread in rail traction until power electronics came along that could supply it with a variable frequency depending on actual and desired speed.

 

[Thebaz]

All this talk about thickness of wires has reminded me to ask why in some situation OHLE has been installed as a rail instead of a wire, for example at the Thameslink platforms at St Pancras International?

 

[swt_passenger]

All this talk about thickness of wires has reminded me to ask why in some situation OHLE has been installed as a rail instead of a wire, for example at the Thameslink platforms at St Pancras International?

At St Pancras it’s to maintain a more accurate position of the catenary on a two dimensional curved section. The rails carry the same contact wire you get elsewhere on the route.

 

[GraphitProjekt]

A universal motor is basically a DC motor, feeding current into stator coils and via a commutator into armature coils so that the magnetic field switches direction to force the armature to move. The coils form an inductor, so will have an impedance that increases at higher AC frequencies. This means that at higher frequencies more of the power is wasted as heat in the coils, not only reducing efficiency but also limiting the motor power according to how well it can be cooled. I guess this was thought to be intolerable at 50Hz or 60Hz grid frequencies so a lower frequency was used (although at that time grid frequencies probably varied quite a bit place to place).

That's interesting! There's also another factor i've heard about which i'll just copy in from Garry Keenors book

"Some administrations experimented with series DC motors but found that when AC currents were
fed through the motor brushes and commutator, the momentary short-circuiting that is inherent to the design
led to induced voltages and high currents. This transformer effect produced significant arcing at the brushes as
they break the current flowing to the armature. The phenomenon, which does not occur with DC current,
significantly increased wear on the commutator. The effect is proportionate to the frequency of the AC
current, so some administrations - notably Germany - adopted a lower OLE frequency," (Keenor 2021: p79)

What's interesting to note about the 15kV 16.7Hz system is that aside from needing a larger step-down transformer on-board the train and relying on electromechanics rather than power electronics, it still pretty much shares all the same advantages as 25kV 50/60Hz over DC systems which is why it was best suited for electrification in the aforementioned countries and why new infrastructure such as the new base tunnels still use it.

Yeah they can run high speed trains up to 300km/h so it has similar operational capabilities to the standard systems. However it does require the railway operator to maintain either a shadow grid system and/or a lot more power electronics to convert incoming 3 phase 50Hz public utility supplies into single phase 16.7Hz power. Germany, Austria and Switzerland actually share a multi-nation spanning transmission network just for traction with it's own power stations distributing at either 110 or 132kV single phase 16.7Hz. From an infrastructure nerd point of view, this is kinda awesome - however I can see why no one is going to start a new system with this technology. If you're railway has been maintaining and operating a transmission system for the last 100 years that's one thing but it's not normally going to worth developing that capability from the ground up.

The 12kV 25Hz system on the North East Corridor in the East Coast of America is pretty unique in that they hang 132kV single phase lines on the same masts as the OLE but several metres higher in the air. I don't know of any other railway that does this.

Indeed low-voltage DC systems typically come along with their own 3 phase AC distribution system as well which are a lot easier to intergrate with existing Grid systems. (Network Rail operate a large network of 11kV, 22kV and 33kV AC distribution lines across southern England fed from the grid at ~35 locations) but if they were originally implemented post-war, this typically isn't the case and each DC substation individually connects to the grid (still trying to found out how it works on Tyne & Wear)

in some situation OHLE has been installed as a rail instead of a wire

I have also heard that "rigid conductor rail" is sometimes used where regular maintenance of the OLE would be especially difficult, I think it's used in Crossrail tunnels and on the Severn Tunnel. Perhaps it's like the slab track of the OLE world. Also, obviously, you'll find it on moving OLE in depots and swing bridges

 

[Bald Rick]

All this talk about thickness of wires has reminded me to ask why in some situation OHLE has been installed as a rail instead of a wire, for example at the Thameslink platforms at St Pancras International?

Principally for maintenance, ie it doesn’t need any other than occasional inspection. Also for reliability. Iver never known a conductor bar OLE system to break.

 

[Rush Matafu]

Welcome to the forum!
Diesel Shunters?
Have a look at the "large shunters" section of this article - should be able to tell you more about the classes of shunters which are still in operation!

List of British Rail modern traction locomotive classes - Wikipedia

en.wikipedia.org

Thank you, it was very helpful!

 

[AlastairFraser]

Thank you, it was very helpful!

No worries!

 

[ac6000cw]

However it does require the railway operator to maintain either a shadow grid system and/or a lot more power electronics to convert incoming 3 phase 50Hz public utility supplies into single phase 16.7Hz power.

You don't need power electronics - historically rotary converters were used to convert 50Hz 3-phase to 16 2/3 Hz single phase, as an alternative to generating it directly in a dedicated power station. That was a major reason 16 2/3 Hz was originally chosen as the frequency i.e. exactly one-third of 50Hz. It was the decreasing (whole life) cost and increasing capability of power electronics that swung the balance that way eventually.

 

[MarkyT]

A universal motor is basically a DC motor, feeding current into stator coils and via a commutator into armature coils so that the magnetic field switches direction to force the armature to move. The coils form an inductor, so will have an impedance that increases at higher AC frequencies. This means that at higher frequencies more of the power is wasted as heat in the coils, not only reducing efficiency but also limiting the motor power according to how well it can be cooled. I guess this was thought to be intolerable at 50Hz or 60Hz grid frequencies so a lower frequency was used (although at that time grid frequencies probably varied quite a bit place to place).

I think there was also heavier sparking at the commutator with higher frequencies, shortening its life.

Grid frequency electrification wasn't an option until the advent of rectifiers suitable for use on board, which would convert the AC into DC to feed it into a DC motor. These were initially mercury arcs but rapidly superseded by semiconductors.

I doubt there was widespread use of mercury arc rectifiers onboard rail vehicles, though they were definitely used in fixed DC substations. It could be very dangerous in an accident for example if several Kg of mercury was spilled on a moving train.

Finally, the rotation speed of an AC motor is closely tied to the supply frequency. This obviously isn't ideal for a train that has to run at a range of speeds, so they didn't become widespread in rail traction until power electronics came along that could supply it with a variable frequency depending on actual and desired speed.

Italy used AC motors for their early 3-phase systems. The traction units had a small range of fixed motor speeds that could be selected.

 

[Energy]

I doubt there was widespread use of mercury arc rectifiers onboard rail vehicles, though they were definitely used in fixed DC substations. It could be very dangerous in an accident for example if several Kg of mercury was spilled on a moving train.

They were... class 86, 87s 81-84 and others had them.

 

[MarkyT]

They were... class 86, 87s and others had them.

Ok. According to this thread: https://www.railforums.co.uk/threads/the-ac-electrics-classes-80-85-86-87.214157/ , it was the earlier AC electric classes 81-84 that had mercury rectification. The rectifiers were very heavy reinforced units, based on early French AC practice, apparently; Classes 85 onwards used germanium or silicon solid state.

 

[Richard Scott]

They were... class 86, 87s and others had them.

No they didn't, these had solid state rectifiers as did 85s, I believe. 81s to 84s all had mercury-arc rectifiers when built but believe most were retrofitted with solid state rectifiers fairly early on. 87101 was also fitted with Thyristor control but think normally it was used as a standard locomotive as needed a key to be inserted to use it in Thyristor mode. There were certainly locos in service with Thyristor control in Germany in 1970s (class 151, think they were built in early 1970s) and probably other classes in other countries?
Adding to 15kV 16.7Hz discussion, sorry if someone already said this, but originally the 16.7Hz was used as could be used with DC motors without rectification. 50Hz would have caused unacceptable arcing on commutators whereas 16.7Hz kept it within an acceptable limit. Obviously with modern AC drives this isn't an issue, just have to put up with bigger transformers, as already mentioned.

 

[ac6000cw]

81s to 84s all had mercury-arc rectifiers when built but believe most were retrofitted with solid state rectifiers fairly early on

That's my understanding too.

AFAIK the Glasgow 'Blue Train' class 303 EMUs also originally used mercury-arc rectifiers, as did the LTS class 302. With their lower power requirements, EMUs transitioned to using solid-state rectifiers quite early in the 1960s.

Re. locos in other countries using mercury-arc rectifiers, in the US two railroads with low-frequency AC electrification bought Ignitron rectifier equipped freight locos from GE starting in 1955 with twelve 3300hp EL-C (Virginian Rly.), followed by sixty 4400hp E44 in 1960 (Pennsylvania RR).

There were certainly locos in service with Thyristor control in Germany in 1970s (class 151, think they were built in early 1970s) and probably other classes in other countries?

The seminal Thyristor controlled loco design is probably the ASEA-designed Swedish Rc series, with the Rc1 version in service from 1967. A notable derivative of the design was the 7000hp 125mph triple-voltage/dual-frequency Amtrak AEM-7 'Mighty Mouse', entering service in 1980 (and a very worthy successor to the classic 1934-vintage GG1).

 

[MarkyT]

AFAIK the Glasgow 'Blue Train' class 303 EMUs also originally used mercury-arc rectifiers, as did the LTS class 302. With their lower power requirements, EMUs transitioned to using solid-state rectifiers quite early in the 1960s.

Makes sense. I guess the solid state tech improved over the years to handle ever higher currents.

Re. locos in other countries using mercury-arc rectifiers, in the US two railroads with low-frequency AC electrification bought Ignitron rectifier equipped freight locos from GE starting in 1955 with twelve 3300hp EL-C (Virginian Rly.), followed by sixty 4400hp E44 in 1960 (Pennsylvania RR).

I'd not heard of 'ignitron' before. Seems it was a patented later variant of the mercury arc system with an alternative method of striking the arc. https://en.wikipedia.org/wiki/Ignitron

The seminal Thyristor controlled loco design is probably the ASEA-designed Swedish Rc series, with the Rc1 version in service from 1967. A notable derivative of the design was the 7000hp 125mph triple-voltage/dual-frequency Amtrak AEM-7 'Mighty Mouse', entering service in 1980 (and a very worthy successor to the classic 1934-vintage GG1).

The Amtrak locos were also known as 'meatballs' and 'swedish toasters'!

 

[Elecman]

They were... class 86, 87s and others had them.

86/87s never had Mercury Arc rectifiers it was the first generation AC electric classes that were fitted with them. the class 85 locos and onwards had diode rectifiers

 

[Energy]

Ok. According to this thread: https://www.railforums.co.uk/threads/the-ac-electrics-classes-80-85-86-87.214157/ , it was the earlier AC electric classes 81-84 that had mercury rectification. The rectifiers were very heavy reinforced units, based on early French AC practice, apparently; Classes 85 onwards used germanium or silicon solid state.

No they didn't, these had solid state rectifiers as did 85s, I believe. 81s to 84s all had mercury-arc rectifiers when built

86/87s never had Mercury Arc rectifiers it was the first generation AC electric classes that were fitted with them. the class 85 locos and onwards had diode rectifiers

Many thanks, everyone - post now corrected!

For those curious, the rectifier from a class 81 (I missremembered it as 87, whoops!) looks like this:


[credit]

 

[JohnElliott]

AFAIK the Glasgow 'Blue Train' class 303 EMUs also originally used mercury-arc rectifiers, as did the LTS class 302. With their lower power requirements, EMUs transitioned to using solid-state rectifiers quite early in the 1960s.

They did originally use mercury-arc rectifiers, but the rectifiers were inclined to backfire, which damaged the transformers and led to fires and explosions. So it's not surprising that they switched to solid-state as soon as they could.