Distributed rectification for third rail

[HSTEd]

I'm not sure how you would get the current from the insulated conductor rail to the overhead wire?

What insulated conductor rail and what overhead wire?

EDIT:

Let's take an example scheme like Bidston-Wrexham. Something like 50km so at the outside 50 of these 300kW substations. Which is a total of 15000kWe.
Apparently the minimum allowed voltage on the third rail in more modern standards is now 500V - so I will use that as it will simplify the calculations.
That means that the absolute maximum current that a substation can provide is now 600A. Which means that the absolute maximum the scheme can produce is something like 30kA. If the currents in the conductor rail exceed 9kA at any substation [using a hall effect sensor on the conductor rail jumpers] an intertrip will be signalled.

That means the largest fault not detected in that manner is 18kA. [Just under 9kA from each direction].

Hall sensors also have the advantage that they allow quantification of the direction and size of the current quite easily - especially when we have microcontrollers at our disposal.
It is entirely possible to set a limit for the current consumption of any one "segment" of the line between two substations, or for any other arbitrarily determined section of the track bounded by the sensor positions.
If that value is exceeded - and for our 4-car train examples I would set that at something like 5000A total (2 1000kW trains at 500V accelerating plus some margin) then an intertrip will trigger.

So now our maximum fault current is 5000A.
This means that 2500A will flow in from each side of a fault, which implies a maximum (500m to substation) voltage drop of 16.25V.
16.25V means that 5000A must flow through a minimum voltage of ~482.5V to avoid an undervolt intertrip.

Which means the minimum allowable fault resistance must be 0.1ohm. Any lower will cause one of the intertripping systems to go off.
Since 100 milliohm is much greater than the loop resistance to the first substation it appears that the majority of the fault power will be pumped directly into whatever it is. Which would be something approaching 2.5MWe into that resistance.

If its a tree branch or some resistance wire or something it will disintegrate in a couple of seconds. Just like tree branches do when they come into contact with power lines.
You could further increase the minimum fault resistance by chosing values for the maximum allowable current before an overcurrent intertrip to be suitable to the situation - for example there is no reason why 9kA should be flowing in the conductor rail on a Bidston-Wrexham scheme, that is merley the maximum allowed by the third rail standard.

 

[hwl]

How would it 'know', given that in normal operation the substation will be running flat out to supply traction current (with the shortfall provided by adjacent substations)?

You mean exactly like the Gatwick Aluminium can... (and plenty of others)

When I originally read the thread (on phone so no immediate response) I saw 4 main issues:
a) Protection
b) lack of sufficient spare captivity on DNO's networks without installing lots of new supplies (especially in the area concerned, i reason for the introduction of smart metering is to reduce the need for DNOs adding capacity to their networks also at time when electric vehicle charge will increase demand on networks - see todays Times for the cost of electric vehicle charging infrastructure due to high DNO costs in come locations...)
c) most modern 4car EMUs are software limited to circa 1650A on 3rd rail which is 25% more than a 319s.
d) it isn't a solution for the busier branches like Uckfield which reduces the market.

a and b have been briefly addressed by elecman and AM9 but there are major issues with achieving the required disconnection times with multiple "pseudo" isolators and communication via SCADA that I don't think you have understood.

 

[hwl]

Because it knows the voltage at which it is outputting current.
If the voltage collapses as a result of something shorting out the rail then it can detect a low-voltage warning and conclude something is wrong.
For example the minimum voltage permitted in normal operation on third rail is 450V (I imagine the voltage would be set rather higher than that in this case, but you get the idea).

Even if the short is sufficiently high resistance to avoid detection by a local under-volt - hall effect sensors on the conductor rail jumpers will be able to detect enormously high currents in the rail. And thats assuming the control room doesn't initiate an intertrip on the basis of every single substation in the isolation area all jumping to max current at once.

EDIT:
For example - a dead short has an absolute minimum impedance equal to the loop of rails between it and the supplying substation(s).
If we assume a project where there are 20 substations at intervals in one direction from the fault (as a fault is symmetrical).
The loop resistance to the first substation (assuming all are operational for the moment) is at most 6.5 milliohm.
That means that in order for a dead short to reach 450V at that substation's busbar and avoid an undervolt (I would set higher, perhaps 600V myself) you would have to have a fault current flowing of nearly 70,000A.
Since all substations would be electronically limited to 300kW and to a current proportional to it then it is highly unlikely they would be able to supply such a current without going into undervolt.

In order to avoid going into an undervolt the resistance of the short (whatever it is) would have to be very much greater than the loop resistance of the conductor rails - which would then mean that effectively all the heating would be passing into whatever it is that is bridging the conductor rails.
For example a one ohm resistance across the rails would have a minimum of 300kW being dispersed in it, probably much more.
Whatever it is will be blown clear rather quickly with minimal damage to the conductor rail or rectifier system - which could quite easily keep that up all day;
It is also highly unlikely that such a fault would trip conventional circuit breakers anyway.

How does it sense the difference between a fault and soggy ground /insulators after raining?

 

[rebmcr]

What insulated conductor rail and what overhead wire?

Can't see that happening, then.

 

[HSTEd]

How does it sense the difference between a fault and soggy ground /insulators after raining?

Well modern polymeric insulators aren't supposed to be that heavily influenced by raining - but unless these insulators are going to cause thousands of amps to flow then it should just ignore it.

The objective of fault protection is not to protect whatever causes the fault - it is to protect the equipment itself. Unless the fault currents are going to cause immediate damage then there is no point isolating. Just raise a flag in the control room.

Can't see that happening, then.

And yet it doesn't appear like much more overhead wiring is going in either

 

[rebmcr]

And yet it doesn't appear like much more overhead wiring is going in either

It's a valid point (although I am a little more optimistic about postponement rather than cancellation) -- but nevertheless, the ground-level conductor is such a deal-breaker that I don't see any scheme including it to be a viable alternative. :/

 

[HSTEd]

It's a valid point (although I am a little more optimistic about postponement rather than cancellation) -- but nevertheless, the ground-level conductor is such a deal-breaker that I don't see any scheme including it to be a viable alternative. :/

That is unfortunatley more because Network Rail has decided that it wants 'proper' electrification and would rather have none than settle for 'second best'.

I've been digging through regulations and in segregated operations there is no prohibition on it. It is more a management culture thing. HMRI or similar would have a very hard time arguing that it shouldn't be allowed.
Especially since it approved major expansions as late as the ELL extension.

 

[Elecman]

That is unfortunatley more because Network Rail has decided that it wants 'proper' electrification and would rather have none than settle for 'second best'.

I've been digging through regulations and in segregated operations there is no prohibition on it. It is more a management culture thing. HMRI or similar would have a very hard time arguing that it shouldn't be allowed.
Especially since it approved major expansions as late as the ELL extension.

NR has to comply with the requirements of the ORR and the ORR has already decreed no more 3rd Rail except for very minor extensions ( new sidings etc)

 

[HSTEd]

NR has to comply with the requirements of the ORR and the ORR has already decreed no more 3rd Rail except for very minor extensions ( new sidings etc)

And because some bureaucrat made this decision for unknown or unsubstantial reasons and because NR is not willing to fight over it for its own selfish reasons (it apparently either prefers diesel operation or thinks it can extort enough money for its preferred electrification scheme from the government) we end up with no electrification of large parts of the network that would otherwise be electrified.

And yet it approves large quantities of new third rail for ELL and in other places.
It would help if it was somewhat consistant.

 

[Elecman]

Pedantic it Approved the installation on the ELL some years ago, it wouldn't now. NR doesn't have a choice, it is publicly funded now via the Treasury so it does as its political paymaster says / funds not what it may or may not want other than via the 5 yearly negotiations via the ORR as part of the CP regime. ( the government sets the funding and NR/ORR agree what can be delivered for that funding. It would seriously help if the ORR stopped setting stupid efficiency targets for each period as you can only pare off so much before it becomes impossible to deliver the other targets. Especially when the ORR then moves the goalposts with regard to Safety regulation mid period/ delivery timescale

 

[dviner]

And because some bureaucrat made this decision for unknown or unsubstantial reasons and because NR is not willing to fight over it for its own selfish reasons (it apparently either prefers diesel operation or thinks it can extort enough money for its preferred electrification scheme from the government) we end up with no electrification of large parts of the network that would otherwise be electrified.

And yet it approves large quantities of new third rail for ELL and in other places.
It would help if it was somewhat consistant.

Why bring the ELL extension into the whole ORR/NR "no more 3rd rail" policy?

 

[HSTEd]

I went digging through the 15/16 RSSB Risk and fatality report for electrocution industries - interestingly the majority (63%) of workforce electric shocks are actually the result of non-traction supplies. I assuem that would be lighting and signalling systems or whatever.

Anyway there was one passenger fatalty as a result of a conductor rail - passenger fell off a platform and came into contact with it. There have been no workforce electrocutions for a while and trespass ones aren't broken down.

If one passenger fatality is enough of a safety case - shouldn't we be banning platform signs because one killed a passenger when it fell off last year.

 

[Elecman]

there has been 2 fatal public falls from platforms onto,the 3rd rail in the last 2 years. ( from the RSSB report

• Since 2011/12, there have been three occasions where a person falling from the platform has subsequently come into contact with the conductor rail. Two of these events, both in the last two years, have been fatal; the likelihood of fatality is comparatively high when this type of accident occurs.

I have not aware of any accidental falls from the publicly accessible infrastructure resulting in Fatal electric shocks concerning the OLE.

 

[HSTEd]

Two fatalities in five years for a thousand route km or more is pretty small.

The 2016 Value for Preventing a Fatality (VPF) is £1,826,000 thanks to the RSSB.
That translates into something like £730k/yr for the entire third rail network.

So about £730/yr/route-km.

At that rate it would take an awful long time for it to overturn the advantage over 25kV at its current prices.

 

[najaB]

Two fatalities in five years for a thousand route km or more is pretty small.

The 2016 Value for Preventing a Fatality (VPF) is £1,826,000 thanks to the RSSB.
That translates into something like £730k/yr for the entire third rail network....

As you're fond of pointing out, fatalities are so uncommon that it's pointless using them as the basis of calculations. This works both ways...

 

[HSTEd]

As you're fond of pointing out, fatalities are so uncommon that it's pointless using them as the basis of calculations. This works both ways...

Indeed, but you can still use it as an order of magnitude estimate.
The value is very small - and the cost justifiable to avoid the deaths is very small. Urgo this is not a reason to ban third rail given the current price of 25kV electrification.

(You can't say that the saving must be enormous because the numbers are too small for a good calculation).
I have put in an FoI for the listed document, or at least the section of it containing the table that is mentioned above. Will get back when they reply.

 

[Elecman]

Not at all one fatality is one too many!!

 

[najaB]

I have put in an FoI for the listed document, or at least the section of it containing the table that is mentioned above. Will get back when they reply.

Please do. Remember though, all it takes is one instance of 'dog runs onto track, little Tommy runs after it and gets killed, along with Sally and Rasheed (who don't know better and try to help)' for the figures to change quite significantly.

 

[HSTEd]

Not at all one fatality is one too many!!

Then lets ban station signs.
And probably trains

And probably ban everything.

 

[dviner]

Not at all one fatality is one too many!!

Going by the RSSB 2015/16 report, one of the eight passenger fatalities may not have occurred if it wasn't for that big lump of conductor rail...

However, we shouldn't be flippant when it comes to fatalities. What is encouraging - reading the RSSB report - is how FEW non-trespass/suicide fatalities there have been in the 2015/2016 reporting period.

This does, however, get away from the main subject of the thread - which I believe was about how great big chunks of metal balanced on porcelain pots could be cheaper than dangling wires from posts.

 

[HSTEd]

Also note that both the passenger fatalities were from passengers falling off a platform and hitting the rail.

This does raise the question as to whether it might not be possible, certainly in this proposal where the train is 80m long at least, to gap the conductor rail in the centre of the platform.

If we know every train using the station is 80m long and has shoes something like 75m apart, then we could put a 60m gap in the platform and then people falling of the platform would be less likely to hit a bit with rail.

 

[najaB]

If we know every train using the station is 80m long and has shoes something like 75m apart, then we could put a 60m gap in the platform and then people falling of the platform would be less likely to hit a bit with rail.

It's a nice idea and probably has merit. But you know Sod's Law will mean that it's a train with a missing shoe that ends up stopping short.

 

[HSTEd]

It's a nice idea and probably has merit. But you know Sod's Law will mean that it's a train with a missing shoe that ends up stopping short.

Or you could leave the rail in place in the centre but have it isolated unless that happens. Motor Switch (or manual for that matter as it would be rarely needed) Disconnector in the substation positioned at the station that energises it.

 

[Domh245]

And because some bureaucrat made this decision for unknown or unsubstantial reasons and because NR is not willing to fight over it for its own selfish reasons (it apparently either prefers diesel operation or thinks it can extort enough money for its preferred electrification scheme from the government) we end up with no electrification of large parts of the network that would otherwise be electrified.

And yet it approves large quantities of new third rail for ELL and in other places.
It would help if it was somewhat consistant.

I may be misremembering, but wasn't one of the take aways from the recent parliamentary commitee into GWEP overuns that the relevant bodies are moving away from "Electrify this" to "deliver such and such improvements" leaving NR with several choices as to what it can do. Presumably then, NR might start fighting for 750V DC where it feels that it is the best value way of delivering the required improvements. Whether or not it would actually be the best value approach, or if the places where it makes sense (NDL and Marshlink) are places where the DfT/ORR ask for improvements remains to be seen.

 

[HSTEd]

Further research indicates that it might be possible to use spinoffs of Solid State Transformer technology to make the 300kW substations capable of regenerating if the voltage of the rail rises above the programmed value (I would imagine 10V or so of hysterisis would be a good idea though to stop slightly different voltage settings causing circulating currents between adjacent substations).

Assuming the grid operator doesn't have a heart attack at the thought of the loads becoming generators that solves the regeneration issue.
But I would do some more reading to see if it really is feasible without putting the converter price through the roof.

 

[QueensCurve]

Forgive me if I haven't followed the engineering details of what is proposed in this thread.

Is it not more or less the system used on the Southern 3rd rail network already with a 33kV grid feeding the 3rd rail at short intervals?

 

[QueensCurve]

The risks of electrification are well known and legislation/codes of practice have been hohned over the years to bring the safety, reliability and capital risks down to politically acceptable levels.
​

Which is why the electrification programme is in total meltdown with massive cost overruns, huge postponements and descoping?
The 25kV programme has failed miserably to deliver on its promises of reduced capital costs and it has probably set electrification back 20 years or more already.

The 25kV project is in meldown because it deviated from the practices honed over 50 year and introduced new standards that failed to take acount of previous practice, knowledge and history.

 

[QueensCurve]

And so electrification is dead.

Not dead, but it may be in a coma for another 15y.

 

[HSTEd]

Forgive me if I haven't followed the engineering details of what is proposed in this thread.

Is it not more or less the system used on the Southern 3rd rail network already with a 33kV grid feeding the 3rd rail at short intervals?

Essentially I am proposing that given the drastic reduction in the size, cost and minimum practical size of rectification equipment - and given the microcontroller control and hall effect current sensors we now have at our disposal, that we should give up on the large substation every few miles concept and have many small substations all working in concert.

Because of the way that I^2R works, it is better to supply current as close as possible to the train - even if you can only provide some of the train's current demand.

The 25kV project is in meldown because it deviated from the practices honed over 50 year and introduced new standards that failed to take acount of previous practice, knowledge and history.

Those old standards cannot be ressurected though as the new safety case environment renders them dead.
We will never be permitted 150mm 25kV clearances again, nor any of the time saving techniques developed by BR in the old days.

Not dead, but it may be in a coma for another 15y.

Which means it will be 30 before any new electrification projects are completed after the ones in the pipeline - which is a disaster.

 

[QueensCurve]

Those old standards cannot be ressurected though as the new safety case environment renders them dead.
We will never be permitted 150mm 25kV clearances again, nor any of the time saving techniques developed by BR in the old days.

A Safety Case is a body of evidence to show that safeguards are in place to prevent reasonably foreseeable hazards and that risks have been reduced to ALARP. A Safety case is provided to facilitate the work activity not to frustrate it.

In this case it might be argued that the electrocution or flashover risks at the clearances used previously are not so big as to be reasonably foreseeable. If the risks are assessed it could be argued that it goes beyond what is reasonably practicable to increase the clearances to the new level.

Of course the Safety Case will have to show compliance with relevant legislation and standards and that is where the issue arises. In this case however the standard provided for a derogation to be available for projects that were in progress.