Overhead Line Power Capacity

[Richard123]

So with this how would it be possible for the first feeder station to feed beyond the midpoint track section to its right?


I've got down a bit of a rabbit hole over the last few days of looking at railways in Europe. I didn't realise how many different voltages are used: 1.5kV DC in France and the Netherlands, 3kV DC in Belgium and 15kV 16 2/3 Hz in Germany and Switzerland. It makes you wonder why all these different voltages were chosen, presumably due to technical limitations many decades ago. The odd German frequency is the most intriguing.

Speaking of which, HS1 and the Channel Tunnel must have a very good power capacity as each Class 374 set can draw up to 16MW of traction power. A couple of those accelerating and the power supply would know about it!

16.7Hz had benefits reducing commutation problems with motors of the time - others on the group may be able to elaborate.

It also greatly helps with voltage regulation, due to lower inductance at that frequency, and all the supplies can run in parallel.

The main disbenefit with hindsight is the selection of 15kV, where a higher voltage would be even more efficient.

There are also slight weight penalties in transformer weight and more complex supply equipment, but these are greatly outweighed by the efficiency benefits of the supply configuration and simpler OLE and distribution compared to
2x25kV systems used in UK/France/Spain for high demand railway.

(West Germany also has its own railway HV grid if you want a new rabbit-hole, a historic arrangement dating to when railways generated their own power before the public grids were established)...

 

[edwin_m]

The odd German frequency is the most intriguing.

16.7Hz had benefits reducing commutation problems with motors of the time - others on the group may be able to elaborate.

It also greatly helps with voltage regulation, due to lower inductance at that frequency, and all the supplies can run in parallel.

The main disbenefit with hindsight is the selection of 15kV, where a higher voltage would be even more efficient.

There are also slight weight penalties in transformer weight and more complex supply equipment, but these are greatly outweighed by the efficiency benefits of the supply configuration and simpler OLE and distribution compared to
2x25kV systems used in UK/France/Spain for high demand railway.

(West Germany also has its own railway HV grid if you want a new rabbit-hole, a historic arrangement dating to when railways generated their own power before the public grids were established)...

An AC system allows a higher voltage to be stepped down by a transformer on the train. Particularly for main line railways the higher voltage reduces energy losses and allows feeders to be much further apart.

As mentioned, motors at the time couldn't really work at the industrial frequency of 50Hz and there was no viable way of turning this into DC on board. So they picked a frequency that would work with the motors they had, and is also exactly one third of 50Hz so can be provided from a rotary converter with on third the number of windings on the output side.

The 25kV system became viable in the 1950s when technology allowed for rectifiers to produce DC on board the train, to be fed into a DC motor. These were initially mercury arc but replaced by semiconductors in the 1960s. Thus the 50Hz system avoided rotary converters and the higher voltage probably reduced losses further. But Germany and nearby countries had so much of their networks electrified by then that it wasn't worth them changing to a new system.

 

[GraphitProjekt]

So with this how would it be possible for the first feeder station to feed beyond the midpoint track section to its right?

Mid point substations have a circuit breaker on the busbar that is normally open, but can be closed if safe to do so. It's probably interlocked with other breakers to make sure different grid feeds never touch.

I've got down a bit of a rabbit hole

I've been burrowing since last october, there's so much fun stuff!

over the last few days of looking at railways in Europe. I didn't realise how many different voltages are used: 1.5kV DC in France and the Netherlands, 3kV DC in Belgium and 15kV 16 2/3 Hz in Germany and Switzerland.

1.5kV on mainline mixed traffic railways is a historical oddity that is now somewhat of a limitation for those countries, especially Netherlands. Japan also has loads of 1.5kV dc. With that being said, it's a modern standard for dedicated metro/suburban systems. There's a suburban line in montreal that recently went from 25kV AC to 1.5kV dc!!

3kV DC is also in Italy, Morroco, South Africa and many places that the USSR had influence.

It makes you wonder why all these different voltages were chosen, presumably due to technical limitations many decades ago.

Back in 1900 there were basically 2 options available to you. LV DC or low frequency AC.

With dc electrification you can generate 3phase ac at (historically) between 6-66kV. This often developed into whole distribution grids fully controlled by the railway and distributed to AC/DC substations for a line voltage of 600 up to 3000. This is fed straight to resistor banks to control dc motors.

Substations spacing is obviously dependent on that line voltage but i think 3000 is the limit possibly due to

1. the size of motor (higher voltage=thicker insulation? Probably other reasons why optimal motor voltage is around 1kV),
2. expense and practicality of dc circuit breaker and rectifier technology (in the early days susbtations were manned with dc rotary converters),
3. and because the longer the length between susbtations the higher the risk that only one of the circuit breakers would actually detect a fault. (In contrast to AC, DC electrification sections are double end fed)
4. Third rail requires lower voltage i assume because of the air gap to the ground. 750V for top contact, but side/bottom contact can get up to 1200V

Substation spacing for 3000V is about 7km i think. For 750V down to around 3 or 4. But of course it depends wildly on the traffic requirements and line speeds etc.

25kV AC at 50Hz is actually quite similar to old school DC. But instead of the train picking up LV DC, it connects to that HV AC distribution network (which in this case is single phase 25kV) and each train is a mobile substation stepping down and rectifying to around 1kV dc for dc motors or (later) a motor drive powering 3phase ac motors. Having a rectifier on a train was only really an option in the 1950s

For this reason, LV DC electrification is still preferred over 25kV AC where you railway is sufficiently short distance and high frequency that if the transformers and rectifiers are in standalone substations rather than on every train - you'll need to buy fewer transformer/rectifiers. A short distance/high frequency railway is otherwise known as a metro.

The low frequency AC thing is to do with 1900s era universal motors being able to work off 16.67 or 25Hz AC rather than 50Hz. Others cleverer than I can understand exactly why this is.

Obviously with AC OLE you can feed the train with a much higher voltage (historically 6.6kV and 11kV) than the motors want because you can step down on board the train, this way minimising substation count. So you can see why some 1900s railways chose this over DC if long distance was important for them.

Both systems often required railways to build power stations and HV distribution grids. (Germany, Austria and Switzerland share a single phase 16.7hz grid just for railways at transmission voltages between 55kV and 132kV). But as public electricity utilities started to appear later on in the 20th century, dc systems could more easily intergrate as they were using 3phase AC sometimes already at 50hz (though 25hz was common with these systems too). For low frequency AC though, this was more of a problem so no one would ever build this system nowadays.

There were a few early railways in the UK that used 25hz 6.6kV. Lancaster-Morecambe-Heysham and (more amusingly) South London under LBSCR: firstly london bridge to victoria but later to crystal palace and some other places. This was ripped up and replaced with 750V 3rd rail in the 1920s under SR to unify the network with the pre-existing 3rd rail out of Waterloo. I'm quite intrigued on the alternate history where southern region is all low frequency AC!

Speaking of which, HS1 and the Channel Tunnel must have a very good power capacity as each Class 374 set can draw up to 16MW of traction power. A couple of those accelerating and the power supply would know about it!

Hs1 is all 2x25kV AT. It also has regular Autotransformer susbtations every 10 or so km to make best use of it.

 

[Tester]

Is this the time to throw in the three phase AC electrification system widely used in Italy up to the 1970s, with double overhead wires (third phase earthed)?

Happy rabbit hole burrowing

 

[GraphitProjekt]

When we're talking about 10MVA, 18MVA or 27MVA for these traction supply transformers, are these figures for summer time or winter time operation? (or in other words, lower or higher capability respectively)

 

[WAO]

A limit on an inductor (like a transformer) is the temperature rise above ambient in the windings. There are a range of classes of insulation which provide for various temperatures which will give a specified service life. Some transformers are naturally cooled, some have forced circulation, which will increase capacity. There is a list somewhere of the respective ratings when natural or forced cooling is applied.

Transformers do take time to warm up (although this may shorten their lives) so can have a useful overload capacity in contrast to a SFC which may need to trip quickly. Protection with thermistor sensors is advisable.

This is a bit vague but may help.

WAO

 

[GraphitProjekt]

Something else I was thinking about recently, how should one go about estimating the distance a feeder station supplies?
I've heard rules of thumb like 18 miles, or 20-30km for the distance between FS and Mid-point. This is clear enough for sections on plain line.

But lets say you had a feeder station at a junction feeding a mainline and a loop-line. The mainline goes out, say, 10 miles and the loop line travels parallel to it for 10 miles as well. So it total, the FS supplies 20 miles. But it occurred to me that if it was just 10 miles of four-track, I would say 10 miles yet there might not actually be a whole lot of difference between these two scenarios.

So instead of counting the route miles, I've been thinking of a different way to count this. What about drawing a circle with the centre over the FS and the circumference at the furthest away neutral section/end-of-wires - and the radius is the actual estimation of how much it feeds. Would this be a more useful figure?

(This was inspired by trying to figure out how the feeding works around Glasgow which is much more dense and spindly than any other area of AC lines, more comparable to South London DC lines)

 

[WAO]

Feeder sites depend on the proximity of the Grid.

Feeding distance depends on traffic density. This has increased markedly above "rules of thumb" from 50 years ago, with both traffic and installed power typically doubling. The system has partly coped due to past conservative design. The types of feed are evolving, with the progression from classic/BT, Autotransformer to SFC enabling longer, higher, more efficient feeding.

Whether a future battery could take the Scotsman to its destination remains to be seen, although I think we are near the asymptote. A clean, light nuclear battery producing electricity directly without needing to boil water for a steam engine would suit nicely.

WAO

 

[GraphitProjekt]

Feeder sites depend on the proximity of the Grid.

Feeding distance depends on traffic density. This has increased markedly above "rules of thumb" from 50 years ago, with both traffic and installed power typically doubling. The system has partly coped due to past conservative design.

Yeah, I've looked into older installations in the Anglia/GN area and it seems like a lot upgrading took place in the late 1990s, sometimes adding new supplies, but also adding new MPTSCs to shorten distances and installing neutral sections at feeder stations which had been built with an extra transformer for failover so that they were both normally working

The types of feed are evolving, with the progression from classic/BT, Autotransformer to SFC enabling longer, higher, more efficient feeding.

I think there can be quite a big discrepancy between the electrical capacity of the supply transformers and the actual allowed firm supply capacity which I imagine is at least partly down to the phase imbalence, so I'd hope to see more SFC upgrades at existing feeder stations to make better use of already existing connections

 

[Nottingham59]

I think there can be quite a big discrepancy between the electrical capacity of the supply transformers and the actual allowed firm supply capacity which I imagine is at least partly down to the phase imbalence, so I'd hope to see more SFC upgrades at existing feeder stations to make better use of already existing connections

Especially if the SFCs can be programmed to take their power from whichever phase will restore the imbalance caused by the existing gird connection.

 

[Richard123]

Especially if the SFCs can be programmed to take their power from whichever phase will restore the imbalance caused by the existing gird connection.

They do improve grid power quality by reducing “background” NPS and harmonics when connected, but grid effects are quite localised so it won’t help another railway site many km away much.

(Also, depending how the grid is configured and varying loads, the existing imbalance itself varies, which means a connection can be compliant one day and non-compliant the next…)

 

[edwin_m]

I don't know for sure but I get the feeling there would be a "natural" imbalance correcting effect if the SFC is just fed by a star-delta rectifier. Whichever phase has a greater magnitude of voltage will feed more current into the DC link - though I guess it might depend how much capacitance there is in the link to smooth out the effect of the inverter drawing current from it.

 

[RichardAsh1981]

I don't know for sure but I get the feeling there would be a "natural" imbalance correcting effect if the SFC is just fed by a star-delta rectifier. Whichever phase has a greater magnitude of voltage will feed more current into the DC link - though I guess it might depend how much capacitance there is in the link to smooth out the effect of the inverter drawing current from it.

True. However operating a star delta rectifier (or any 6 / 12 pulse rectifier) on an unbalanced voltage like this causes uneven stressing of the diodes (which are designed for balanced operation) and may lead to premature failure. This is an issue for Variable Speed Drives which are essentially the same thing as a SFC except with three phase output.

I would expect (but don't know) that the harmonic emissions from even a 12 pulse rectifier would be too great to tolerate at the power capacities involved for rail traction (unless the connection were taken at a higher voltage level, and so lower supply impedance than otherwise required - which is where we started!). I am expecting practical SFC equipment to have what is often called Active Power Factor Correction, which means that each rectifier diode is replaced by an solid state switch which can be switched very rapidly (thousands of times per 50 Hz cycle) to control the flow of energy. The basic use is to force the input current to be a sine wave (and one in phase with the voltage waveform) but as this is all controlled by firmware, lots of other "cleaning up" effects can be engineered in, as Richard123 pointed out.

NPS here is Negative Phase Sequence voltages, which put simply mean that an unbalanced load (like the railway without an SFC) superimposes on the grid a voltage difference which tries to turn other users three phase induction motors backwards. Because the voltage is small, it doesn't succeed, but the internal "fight" in the motor predictably causes it to use more power (paid for) and gets the motor hotter (shortening its lifespan). Understandably customera don't like this, which is why limits are set for Negative Sequence Voltages in the public supply grid.