Cost of Third Rail Electrification

[HSTEd]

I have been trying to find some actual supporting documentation for the estimate given by Network Rail for electrifying the Borderlands line at DC (£207m) however the number just seems to appear in the Merseyside RUS saying analysis has shown it.
But there appears to be no actual documentation of how the estimate was derived.

It also seems to be drastically out of step with estimates from Delta Rail developed as part of their low cost branch electrification work from a couple of years back.

So does anyone have information?

I have put in a request to their transparency system but since they are not bound to respond I doubt they will actually reply as it appears these estimates have been deliberately blown up.

 

[Bald Rick]

I expect you will get a reply, but it might take a few weeks.

I don't know the line personally, but if there are any steel sleepers down there, the track will have to be renewed, and that can be a significant cost.

 

[yorksrob]

Excellent going for looking into it.

Having been fortunate enough to have lived through a period of extensive third rail electrification during the late 1980's and early 90's, I find it peculiar that the last couple of islands on the Southern Region still haven't been done in 2014!

 

[edwin_m]

There's a letter in Modern Railways this month from an engineer with one of the contractors, pointing out the difficulty of attaching pots to concrete sleepers. Presumably this only applies if nobody has had the foresight to lay the sleepers that come pre-drilled for them (and these often appear on track many miles from any potential third rail!). Can any members on the Uckfield, Marshlink or Borderlands lines report any sightings of concrete sleepers and whether they are pre-drilled?

The letter suggests some kind of mounting to the rail fixings instead, as used with TPWS loops. I guess a clamp round the sleeper might be another option.

 

[martynbristow]

Didn't chiltern request network rail to build a new station but it was more expensive than them doing it themself.
I'm not sure if network rail is value for money, there's no incentive to be :/

 

[Bald Rick]

Didn't chiltern request network rail to build a new station but it was more expensive than them doing it themself.
I'm not sure if network rail is value for money, there's no incentive to be :/

That was a long time ago. More recently, various organisations have started out doing projects themselves, and then asked NR to complete them as it would be cheaper. Most notably the Scottish Government with both Airdrie - Bathgate and the Waverley line, and as we all know the Scots are most parsimonious with their pennies.

 

[SpacePhoenix]

Having had a quick read about the Borderlands Line on wikipedia I'm surprised that they're considering going for 3rd rail. I would have expected them to go for overhead with dual voltage stock if it'll be going into 3rd rail teritory

 

[snowball]

I seem to remember reading that steel sleepers were the reason for a high estimate for third-railing this line several years ago.

 

[HSTEd]

Out of interest - why aren't all concrete sleepers delivered pre-drilled?
Is the cost of the drilling operation really significant compared to the cost of maintaining stocks of different sleepers.

Additionally - when were the sleepers last changed on this route? How often are sleepers changed?

 

[Tremzinho]

Actually, Network Rail unilaterally decided to install steel sleepers without consulting any of the local stakeholders, simply because it was the cheapest way to relay the track. I asked a Network Rail official why they did this when they knew electrification was in the Merseyside Rail strategy. The excuse was that because the strategy was just an "aspiration document" they didn't need to consult anyone, and they certainly were not going to spend any extra money on the off chance that electrification would happen in the near future.

The quadrupling of the cost completely destroyed the BCR for the project, and it was pushed to the backburner for the forseable future.

 

[HSTEd]

So I was correct in my assumption - Network Rail deliberately installed steel sleepers to ensure that third rail electrification could not take place?
How long ago was this?
The project will get cheaper and cheaper as time goes by as the date for scheduled sleeper replacement closes in.

 

[yorksrob]

There's a letter in Modern Railways this month from an engineer with one of the contractors, pointing out the difficulty of attaching pots to concrete sleepers. Presumably this only applies if nobody has had the foresight to lay the sleepers that come pre-drilled for them (and these often appear on track many miles from any potential third rail!). Can any members on the Uckfield, Marshlink or Borderlands lines report any sightings of concrete sleepers and whether they are pre-drilled?

The letter suggests some kind of mounting to the rail fixings instead, as used with TPWS loops. I guess a clamp round the sleeper might be another option.

I assumed they were the same concrete "Dow Mac" sleepers used on the Marshlink as for the rest of Kent. Would be surprised if they had used different lot (purely from platform observation - the engineers of the area would likely know more).

 

[HSTEd]

Additionally that old Delta Rail report seemed to indicate that if you fitted insulated chairs (which are available) you can use DC electrification with steel sleepers, changing the chairs might be cheaper than re-sleepering so I will look into it. Although based on the costs in that report the cost of resleepering with concrete would only be £43m, so I doubt that is the only reason for the massive cost overrun.

 

[SpacePhoenix]

Additionally that old Delta Rail report seemed to indicate that if you fitted insulated chairs (which are available) you can use DC electrification with steel sleepers, changing the chairs might be cheaper than re-sleepering so I will look into it. Although based on the costs in that report the cost of resleepering with concrete would only be £43m, so I doubt that is the only reason for the massive cost overrun.

How much would it cost to electrify the same route with overhead?

 

[LNW-GW Joint]

So I was correct in my assumption - Network Rail deliberately installed steel sleepers to ensure that third rail electrification could not take place?
How long ago was this?
The project will get cheaper and cheaper as time goes by as the date for scheduled sleeper replacement closes in.

Less than 5 years ago.
The BCR would still be poor even if it had not been re-sleepered.
There are many more lines higher up the "do next" list, including all the other lines radiating from Chester.

 

[UrbanWorld]

I have been trying to find some actual supporting documentation for the estimate given by Network Rail for electrifying the Borderlands line at DC (£207m) however the number just seems to appear in the Merseyside RUS saying analysis has shown it.
But there appears to be no actual documentation of how the estimate was derived.

It also seems to be drastically out of step with estimates from Delta Rail developed as part of their low cost branch electrification work from a couple of years back.

I know Merseytravel estimated far less than Network Rail. The conclusion was that wires were cheaper than third rail. This baffles me as all the electric infrastructure is about the same in volume and cost with the difference between the cost of laying third rail and building 100s or 1000s of overhead wire pylons. 3rd rail to me "appears" cheaper. Merseytravel and Network Rail were in disagreement regarding the Borderlands Line. Network Rail shunned the line.

Wiki has a bit on it and has links.
http://en.wikipedia.org/wiki/Borderlands_Line

"The most recent study, conducted by Network Rail in 2008, investigated the costs of extending the Merseyrail network third-rail electrification to Wrexham. However, when the cost was estimated at £207 million, Merseytravel stated that cheaper overhead-line electrification would also be considered and announced a lower estimated figure of £66 million. This would require dual-voltage trains with third-rail and overhead-wire capability"​

There is a massive difference between £66 million and £207 million. Merseyrail is acquiring 3rd rail and overhead wire trains from Thameslink.

 

[HSTEd]

Less than 5 years ago.
The BCR would still be poor even if it had not been re-sleepered.
There are many more lines higher up the "do next" list, including all the other lines radiating from Chester.

I am not sure the lines radiating from Chester are really suitable for third rail electrification - and thus the line is not competing for the same industrial resources as those lines. (Since I am told continuously that the limiting factor on OLE installation is not how much money is available).

It is worth noting that it is likely that third rial would be significantly cheaper than the OLE if not for the sleepering - which was likely done by Network Rail as part of its 25kV or nothing agenda.

 

[UrbanWorld]

This link shows the spat between Merseytravel and Network Rail.

http://www.penmorfa.com/Wrexham/electrify.html
--- old post above --- --- new post below ---

Excellent going for looking into it.

Having been fortunate enough to have lived through a period of extensive third rail electrification during the late 1980's and early 90's, I find it peculiar that the last couple of islands on the Southern Region still haven't been done in 2014!

Third rail is only extensively in the south east and the Liverpool City Region, with one in Glasgow's Subway and a short line from Manchester to Bury which was taken up about 25 years ago. None in Yorkshire. Any more about?

 

[edwin_m]

There is a massive difference between £66 million and £207 million. Merseyrail is acquiring 3rd rail and overhead wire trains from Thameslink.

Northern (which is a different franchise from Merseyrail) is acquiring some 3rd rail and overhead wire trains from Thameslink, but the 3rd rail shoes are being removed before they go into service in the North, as they are non-retractable and would probably strike various bits of ballast if left in place. The trains are probably also unsuitable for the steep gradients in the Loop line tunnels, and may be too large to fit anyway.

Cost of electrification isn't just what you can see. Third rail requires substations every few miles and often a high-voltage trackside feeder to power the substations. OLE needs very little apart from the OLE itself and the grid supply points which are at much wider spacings. Once in service third rail uses about 25% more energy and trains have lower performance. These factors are worst for longer-distance and rural routes. And rebuilding of bridges for OLE normally also provides clearances for larger containers.

 

[DownSouth]

Out of interest - why aren't all concrete sleepers delivered pre-drilled?
Is the cost of the drilling operation really significant compared to the cost of maintaining stocks of different sleepers.

Would there need to be 'drilling' or just a bung added to the mould so the hole is created when pouring the concrete?

All new concrete sleepers delivered in South Australia or Victoria have one side fitted with chairs for two rails to allow them to be used for either 1435mm or 1600mm lines - even though it is different infrastructure operators running the different gauge networks with most lines having little chance of future conversion! If it is economic for Australian sleeper manufacturers to add extra clips to hundreds of thousands of sleepers to avoid having two separate production runs, it would surely be even simpler for all new UK concrete sleepers to have the holes for third rail mountings included when pouring the concrete.

 

[HSTEd]

Cost of electrification isn't just what you can see. Third rail requires substations every few miles and often a high-voltage trackside feeder to power the substations. OLE needs very little apart from the OLE itself and the grid supply points which are at much wider spacings. Once in service third rail uses about 25% more energy and trains have lower performance. These factors are worst for longer-distance and rural routes. And rebuilding of bridges for OLE normally also provides clearances for larger containers.

Rural routes do not often have significant freight traffic, with some minor exceptions.
The third rail energy losses assume southern railway main line traffic densities and do not apply in a rural setting - as there will be far lower currents in the conductor rail, holding down the voltage drops, especially if aluminium/steel composite conductor rails are used.

There has even been talk of using massed low power ~500kW substations in parallel since they can make use of 400Vac circuits instead of 11kV systems and end up far cheaper per kW of available power.

Additionally the density of substations to support a low density third rail route would be far lower than you suggest, and in the case of the Borderland lines could be positioned entirely at stations which will have handy power supplies as a matter of course.
There is also the fact that you have to use at least a 132kV feeder to support a 25kV network because of the horrific nature of the load - horribly distorted, single phase and hugely peaking.
In some rural areas this would be problematic.

 

[LNW-GW Joint]

I am not sure the lines radiating from Chester are really suitable for third rail electrification - and thus the line is not competing for the same industrial resources as those lines. (Since I am told continuously that the limiting factor on OLE installation is not how much money is available).

It is worth noting that it is likely that third rial would be significantly cheaper than the OLE if not for the sleepering - which was likely done by Network Rail as part of its 25kV or nothing agenda.

I wasn't thinking about 3rd rail, just 25kV.

Both sides of the cost argument are discredited.
Merseytravel said all sorts of things when it was keen to take over its infrastructure from Network Rail, but once they discovered the true cost of maintaining their network they turned tail and fled.
The Network Rail £207m seems preposterous, but every NR electrification scheme is running well over budget, so who do you trust?

Another yardstick is the recent estimate for Oxenholme-Windermere wiring of £16m - just £1m per stkm.
Wrexham-Bidston is 90 stkm (27.5 miles of double track) so £100m seems about the ball park.
The line could also be singled over much of its length, so maybe £60m would do for what is needed.
A dynamic loop in each half of the line, and double track around Dee Marsh for the freight, is all that is needed.

I just think that if you are going to spend anything like £100m on the Chester/Wirral/Wrexham local infrastructure, there are much better things to spend the money on.
I would wire Chester-Crewe/Warrington/Wrexham General with it, plus Halton Curve.

 

[HSTEd]

NR understates the cost of 25kV schemes.

But NR doesn't want third rail any more.

 

[edwin_m]

Rural routes do not often have significant freight traffic, with some minor exceptions.

Maybe so for the Uckfield line, but there's been quite a lot of interest in multimodal freight terminals at Shotton.

The third rail energy losses assume southern railway main line traffic densities and do not apply in a rural setting - as there will be far lower currents in the conductor rail, holding down the voltage drops, especially if aluminium/steel composite conductor rails are used.

But if there are fewer trains there is also less opportunity for regeneration so more power use per train. If the substations are further apart there will be more resistance losses because of greater average distance from the substation. And if the substation distance and rail resistance remain the same, the loss of power is directly proportional to the current so the percentage loss which I quoted does not depend on the power demand.

There has even been talk of using massed low power ~500kW substations in parallel since they can make use of 400Vac circuits instead of 11kV systems and end up far cheaper per kW of available power.

[snip]

Additionally the density of substations to support a low density third rail route would be far lower than you suggest, and in the case of the Borderland lines could be positioned entirely at stations which will have handy power supplies as a matter of course.

A mains supply that can run a few lights and PIDs isn't going to be suitable for traction power, and unless a higher voltage supply is close by the voltage drop in the 400V has to be added to that in the third rail.

Additionally the density of substations to support a low density third rail route would be far lower than you suggest

What would that do for power loss, resilience, and fault detection?

There is also the fact that you have to use at least a 132kV feeder to support a 25kV network because of the horrific nature of the load - horribly distorted, single phase and hugely peaking.
In some rural areas this would be problematic.

It's only needed every few tens of miles, and if necessary inverters can be used to balance the load. A supply covering a large area will be feeding several trains with less peaking (relatively speaking) than a DC substation which effectively only feeds a couple of trains once the resistance of the third rail is taken into account.

 

[HSTEd]

But if there are fewer trains there is also less opportunity for regeneration so more power use per train.

This is not the 50s, a substation can quite easily be equipped with a grid tie inverter pack.
Or in the relatively high voltage-headroom scenario (750V substation voltage I used for initial calculations gives you a full 150V of headroom according to the standard) we have available the overvoltage will be able to drive significant regeneration currents over multiple substation distances, allowing the line to capture most regeneration energy even if you only equipment one substation with regeneration equipment to save money.

If the substations are further apart there will be more resistance losses because of greater average distance from the substation. And if the substation distance and rail resistance remain the same, the loss of power is directly proportional to the current so the percentage loss which I quoted does not depend on the power demand.

The loss of power is proportional to the square of the current.
P = I^2 x R remember.
While the total power that enters the conductor is dependant on the current linearly, so the proportion of the losses increases.

If I have a one amp, ~100V load that I have to drive through a one ohm resistance I will drop 1V and only 99V will drop over the load. 1% was lost.
But if I have a two amp, ~100V load that I have to drive through the same resistance I will drop 2V and only 98V will drop over the load. 2% was lost.
The percentage lost will increase linearly to the average current.

A mains supply that can run a few lights and PIDs isn't going to be suitable for traction power, and unless a higher voltage supply is close by the voltage drop in the 400V has to be added to that in the third rail.

The big big saving in using the 400V supply is that it drastically relaxes the earthing and safety requirements as you no longer need a giant earthing mat or similar. You use simple industrial switchgear rather than fancy HV stuff.
This has been done on various tram projects.

400V supplies are available that can support several hundred amps, and since it would feed an on site rectiformer the exact voltage of the input is irrelevant. Since the voltage will have to be retained inside the norms for an electricity supply and you will be charged on that basis.

Even a standard rural three phase supply will be capable of something like 375A.

What would that do for power loss, resilience, and fault detection?

Thanks to reduced currents the power loss will be less than in an urban setting - and the low currents will still permit resilience.
Fault detection is problematic in all systems like this - having fewer substations doesn't really hurt much since the limiting factor will be the breakers in the paralleling huts.

It's only needed every few tens of miles, and if necessary inverters can be used to balance the load. A supply covering a large area will be feeding several trains with less peaking (relatively speaking) than a DC substation which effectively only feeds a couple of trains once the resistance of the third rail is taken into account.

The supply will still have to be sized for the worst case of all trains at maximum throttle at once - the system is sufficiently small that I can think of several conditions that will stop all the trains at once. (Signal failure for example).
Additionally the DC substations will be spread over a wider series of electrical connections that will make the peaking less of a problem. The substations could be spread over 20 miles in a Borderlands route.
You will load the entire line with multiple ~1600kW trainset loads onto one point.
That will cause all sorts of problems.
And I have considered the whole three-phase rectifier attached to single phase inverter concept for rural 25kV substations - it unfortunately has all sorts of horrific problems that renders it impractical.
You end up with giant capacitor banks otherwise you will get problems with only really loading two of the phases properly, and lots and lots of harmonic distortions. Improvements in power electronics will help (SiC MOSFETs for example) - but that also helps DC systems by allowing simpler synchronous inverter-rectifier topologies and removing the need for big line frequency transformers at substations.
It also allows for such craziness as variable voltage DC substations - so that when a train is drawing power you keep the voltage at the top of the range, but when the line voltage rises as a result of regeneration the substation reduces its holding voltage to provide the headroom.
Significant savings will be available doing that - since the normal substation voltage can increase to ~900V.

 

[Bald Rick]

So I was correct in my assumption - Network Rail deliberately installed steel sleepers to ensure that third rail electrification could not take place?
How long ago was this?
The project will get cheaper and cheaper as time goes by as the date for scheduled sleeper replacement closes in.

That's a rash leap of logic.

Having personally stuck steel sleepers into routes that conceivably, one day, might see third rail electrification I can sure you that the decision to do so was not to 'deliberately... ensure that third rail electrification could not take place'

The decision was taken on the basis that at that time, there was no realistic prospect of third rail electrification arriving in the foreseeable future*. The track needed renewing now, and the cost of doing it with steel sleepers was less than half the cost of concrete - I'm talking millions of pounds cheaper. There was no way the 'funders and regulators' would sanction spending extra millions on something there was no realistic chance of using. And so it has turned out, at least for the 10 years so far.

*oil prices low, serious consideration being given to removing OLE up north for reliability reasons.

On another point, steel and concrete sleepers have a nominal 40 year life.

 

[edwin_m]

400V supplies are available that can support several hundred amps, and since it would feed an on site rectiformer the exact voltage of the input is irrelevant. Since the voltage will have to be retained inside the norms for an electricity supply and you will be charged on that basis.

Even a standard rural three phase supply will be capable of something like 375A.

Thanks to reduced currents the power loss will be less than in an urban setting - and the low currents will still permit resilience.
Fault detection is problematic in all systems like this - having fewer substations doesn't really hurt much since the limiting factor will be the breakers in the paralleling huts.

....

You will load the entire line with multiple ~1600kW trainset loads onto one point.

1600kW at 750V is over 2000A by my reckoning - how does that square with a 375A supply at 400V?

Sorry about the I2R though - I shouldn't have replied after a pint!

You've obviously done some thinking about this and it's a bit time-consuming to find you have an answer for each question but you didn't mention it last time round. Perhaps put it all together in a paper that tells the whole story - but maybe you've done that already but can't post it here?

 

[HSTEd]

1600kW at 750V is over 2000A by my reckoning - how does that square with a 375A supply at 400V?

I believe the idea is that you wire the substations such that they will allow their output voltage will drop if a load tries to draw more than rated current from them.
That will cause power to flow in along the conductor rail from the next substations along and so on until the power actually reaching the train meats demand.

You've obviously done some thinking about this and it's a bit time-consuming to find you have an answer for each question but you didn't mention it last time round. Perhaps put it all together in a paper that tells the whole story - but maybe you've done that already but can't post it here?

Unfortunately I am coming up to my January University exams, so I am trying to avoid spending the time necessary to put these things together. I have not done that much thinking on these topics but I just remember things that I read in other contexts.
After my exams are complete I think I will put something together.

 

[apk55]

With HV AC you could get away with a single substation in the middle at Shotton. There would be no problem in getting power as there is a major switching station for the national grid nearby.
With DC you would probaby need around 8 substations. Finding sites for them could be a problem because as well as needing land you would need road access etc. Then you woud need power lines from bulk grid points as the local grid in many places.

 

[HSTEd]

Eight substations on 27 miles? This isn't the Brighton Main Line.

I haven't done the exact calculations but if you assume the worst case is 4 1600kW units at once then you can probably get by with 4 substations. Efficiency in the worst case scenario is not that important if the system still functions, since the power demand thus voltage drop in the rail under normal circumstances would be fa lower.
I will have done some sums on that though.

Either way the cost of 4 11kV/750V substations will probably come out less than a 132kV/25kV substation which is what you need for HV AC.