Recharging battery trains from 25kV overhead line

[Nottingham59]

Most battery EMUs and tri-mode trains are recharged from 25kV overhead lines. What needs to be considered when designing OHLE in areas where battery EMUs are likely to be used in the future (which probably means all areas)? Particular questions I have include:

• What currents do BEMUs take when recharging? I know the Stadler FLIRT Akku is supposed to recharge in just 15 minutes on the wires; what does that mean in Amps? How many kW did the 802207 trial draw during recharge?
• What is the experience from abroad (and now in the UK) with BEMU recharging? Will any modifications be needed to accommodate the LNER and GC trimode orders?
• EMUs only take a high current when on the move. What changes, if any, are needed to OHLE to accommodate a stationary load of several Megawatts?
• Should BEMUs be designed to use more than one pantograph when stationary to spread the load across several contact points? How should BEMUs deal with poor contact situations?
• How are/should BEMUs recharging be managed to maximise recharge rates without overloading the supply? Is anyone developing standards in this area?
• Does the design of new OHLE schemes like TRU and MML already include provision for future BEMU recharge loads?

Please note that this thread is for 25kV AC and excludes third rail and fast-charging systems like that used by GWR in the Greenford branch trial.

 

[HSTEd]

• EMUs only take a high current when on the move. What changes, if any, are needed to OHLE to accommodate a stationary load of several Megawatts?
• Should BEMUs be designed to use more than one pantograph when stationary to spread the load across several contact points? How should BEMUs deal with poor contact situations?

The static pantograph charging current in the relevant standards is apparently 80A, or a nominal 2MW per pantograph.
Obviously there will be no standing wave effects in the catenary if the train is stationary, so conceptually you could raise any and all pantographs present on the train.

I think a battery train is probably going to operate very differently to a conventional EMU. Transformers are very heavy and have a weight determined almost entirely by rating. I would expect a BEMU to be fitted with a comparatively small transformer and use its batteries to allow for acceleration, even under electric power.

From the perspective of the supply they would likely present an almost constant load. The charger could also be made to respond to voltage of the supply, reducing as the voltage falls, but I am not aware of specific standards.

 

[Nottingham59]

The static pantograph charging current in the relevant standards is apparently 80A, or a nominal 2MW per pantograph.
Obviously there will be no standing wave effects in the catenary if the train is stationary, so conceptually you could raise any and all pantographs present on the train.

80A at 25kV is 2MW per pantograph. So multiple pantograph capability looks like it might be useful.

802207 had, I believe, a battery capacity of around 550kWh, which would give a recharge time of 15 minutes, comparable to the FLIRT akku. Does anyone know how fast the TfW trimodes recharge?

 

[SynthD]

Could the grid and trains communicate frequently, to adjust to the local capacity? If a regular EMU is leaving a station nearby, tell the BEMU to not pull power for the next minute.

 

[Nottingham59]

Could the grid and trains communicate frequently, to adjust to the local capacity?

It would make sense to have some sort of adaptive charging protocol, wouldn't it? Along the line of "If the line voltage drops below 23kV, then only draw as much power as necessary to recharge fully before next scheduled pan-down". "Below 21kV, only recharge enough to get to the next electrification island". "Above 24kV, recharge at the standard rate."

If a regular EMU is leaving a station nearby, tell the BEMU to not pull power for the next minute.

I understand that used to happen (manually) at St Pancras when a Eurostar is pulling out.

 

[Technologist]

The static pantograph charging current in the relevant standards is apparently 80A, or a nominal 2MW per pantograph.
Obviously there will be no standing wave effects in the catenary if the train is stationary, so conceptually you could raise any and all pantographs present on the train.

I think a battery train is probably going to operate very differently to a conventional EMU. Transformers are very heavy and have a weight determined almost entirely by rating. I would expect a BEMU to be fitted with a comparatively small transformer and use its batteries to allow for acceleration, even under electric power.

From the perspective of the supply they would likely present an almost constant load. The charger could also be made to respond to voltage of the supply, reducing as the voltage falls, but I am not aware of specific standards.

To understand the issue here we have a relatively poor connection at the point where the pantograph head and the wire touch with both occasional arching and small surface areas in contact with each other. When we are moving the area of the wire/shoe in contact changes and can thus cool down, when we are stationary current needs to be limited to prevent the wire getting hot and welding itself to the pantograph.

Thus if we wanted to run higher loads we just need to increase the surface area, we could have a stationary pantograph which has an arbitrary number of contact points and thus allow us to pull whatever current the OHLE can take. For a battery train we almost certainly want to be able to stow pantographs into fairings as they cause a surprising amount of drag, which also means we probably want to limit their number too, hence you would have a running unit and a stationary charger unit.

Regarding transformers, you can buy DC-DC converters with 35KW/kg so a 10MW transformer replacement that weighs 300kg is possible. A quick scan of papers shows that electric aviation is targeting ~100KW/kg converters.

 

[HSTEd]

Regarding transformers, you can buy DC-DC converters with 35KW/kg so a 10MW transformer replacement that weighs 300kg is possible. A quick scan of papers shows that electric aviation is targeting ~100KW/kg converters.

Yes, but I'm not aware of anyone having actually built one that light for 25kV single phase.

Even the power electronic converters proposed by ABB and others are still substantially heavier than 35kW/kg.

 

[bahnause]

Could the grid and trains communicate frequently, to adjust to the local capacity? If a regular EMU is leaving a station nearby, tell the BEMU to not pull power for the next minute.

Is a new system really necessary? Usually, regulations on network access specify how railway vehicles should behave if there is a deviation from the 'normal' frequency or voltage.

 

[Nottingham59]

Is a new system really necessary? Usually, regulations on network access specify how railway vehicles should behave if there is a deviation from the 'normal' frequency or voltage.

Sure, but a battery gives the overall "electrification system" much more flexibility to adapt to external circumstances like a voltage drop from other traffic drawing power or the loss of a feeder station. In principle such flexibility could, for instance, allow huge capital savings by needing less redunancy in the grid supply

So there is scope to think about how BEMUs should behave differently from a straight EMUs. I'm not aware of anyone in the UK thinking through the implications, and certainly not seeing any changes coming through in OHLE design standards yet. I don't know what happens abroad, where battery trains are more commonplace.

 

[Technologist]

Yes, but I'm not aware of anyone having actually built one that light for 25kV single phase.

Even the power electronic converters proposed by ABB and others are still substantially heavier than 35kW/kg.

I suspect that's a market issue rather than a technology issue. Given some of the comments around a shortage of transformers for the upcoming needs of AI and cloud computing it wouldn't surprise me if AI, data centres and grid batteries start shifting to power electronics over transformers in those applications before rail.

 

[zwk500]

Sure, but a battery gives the overall "electrification system" much more flexibility to adapt to external circumstances like a voltage drop from other traffic drawing power or the loss of a feeder station. In principle such flexibility could, for instance, allow huge capital savings by needing less redunancy in the grid supply

The most likely thing will be for a train to have an onboard power management system that can adaptively tell the train where to take or send the power from a combination of the pantograph, batteries and regen brakes based on a set of parameters and geofencing.

E.g. It'll only use the batteries if it's got >X% charge left to allow for emergency recovery if needed, it'll only charge the batteries from the Pantograph if it's got a detected voltage within the nominal range. Geofencing could be used to manage regen braking sending power back into the grid and when to charge, as NR could declare an area unsuitable for charging if there's high loads or something.

So there is scope to think about how BEMUs should behave differently from a straight EMUs. I'm not aware of anyone in the UK thinking through the implications, and certainly not seeing any changes coming through in OHLE design standards yet. I don't know what happens abroad, where battery trains are more commonplace.

These discussions are 100% happening in the UK, not least because there's already a system explicitly designed for BEMU charging off the wires being delivered (Cardiff Valley Lines). I was involved in a project in 3-4 years ago that was talking about geofencing charging areas for BEMUs to reopen a freight line to passengers.

 

[Elecman]

You could also increase the number of contact wires above the train at the agreed charging locations

 

[Brubulus]

You could also increase the number of contact wires above the train at the agreed charging locations

This is probably a good way to do it at stations

 

[Nottingham59]

This is probably a good way to do it at stations

Or a thicker conductor bar. If there were multiple contact wires, then i think they would have to be independently sprung so that each wire made a good contact with the Pantograph

 

[ac6000cw]

Or a thicker conductor bar. If there were multiple contact wires, then i think they would have to be independently sprung so that each wire made a good contact with the Pantograph

The OHLE system is designed so that the pantograph lifts the contact wire (and the wire deliberately sags between supports, as it was found a long time ago that it improves current collection). So in reality I suspect getting a reasonably even distribution of contact force between multiple wires wouldn't be a problem. Also 1500V DC systems often use two side-by-side contact wires just to get enough contact area to handle the very much higher currents on that system.

To understand the issue here we have a relatively poor connection at the point where the pantograph head and the wire touch with both occasional arching and small surface areas in contact with each other. When we are moving the area of the wire/shoe in contact changes and can thus cool down, when we are stationary current needs to be limited to prevent the wire getting hot and welding itself to the pantograph.

I guess you could direct compressed air jets towards the contact areas to cool it down (and use temperature sensing/thermal imaging to detect overheat situations), have an additional 'stationary use high current' pantograph with large contact area carbon strips or, as HSTEd suggested, just have multiple pantographs raised to spread the current.

 

[edwin_m]

Or a thicker conductor bar. If there were multiple contact wires, then i think they would have to be independently sprung so that each wire made a good contact with the Pantograph

The conductor bar actually has a copper wire set into the bottom, so the contact area is the same as a normal wire. With the pantograph in contact with one face of a rectangular bar, the skin effect would concentrate the current on the edges of the contact area so the extra width wouldn't help much.

 

[Nottingham59]

Also 1500V DC systems often use two side-by-side contact wires just to get enough contact area to handle the very much higher currents on that system.

That sounds reasonable. Whatever problems there may be getting a high recharge current into stationary trains on 25kV AC, they will almost certainly have been addressed with 1500V DC systems.

 

[Nicholas Lewis]

The conductor bar actually has a copper wire set into the bottom, so the contact area is the same as a normal wire. With the pantograph in contact with one face of a rectangular bar, the skin effect would concentrate the current on the edges of the contact area so the extra width wouldn't help much.

RSSB did some limited research into the current levels that could be transferred under project T1185 (Current Limit at Standstill). The testing demonstrated that a conductor beam with a 120 sqmm copper wire could allow as much as three times higher power transfer than standard contact wire over a 10 minute period without exceeding the temperature limits for the pantograph carbons or conductor wire.

Surprisingly a twin wire setup didn't achieve double but it still facilitates a higher power transfer.

Ultimately the rating of the transformer will determine how much power can be transferred.

https://www.rssb.co.uk/research-catalogue/CatalogueItem/T1185

 

[MarkyT]

The most likely thing will be for a train to have an onboard power management system that can adaptively tell the train where to take or send the power from a combination of the pantograph, batteries and regen brakes based on a set of parameters and geofencing.

E.g. It'll only use the batteries if it's got >X% charge left to allow for emergency recovery if needed, it'll only charge the batteries from the Pantograph if it's got a detected voltage within the nominal range. Geofencing could be used to manage regen braking sending power back into the grid and when to charge, as NR could declare an area unsuitable for charging if there's high loads or something.

I expect a charging plan could be derived for each particular diagram. How much to leave the depot with, where to charge, by how much, contingencies if sufficient supply not available etc. Overall general concerns might include retaining sufficient charge to reach the next but one charging opportunity so a failure of one island can't immediately disrupt a trip, but might trigger a set swap tactic later, or a higher charge rate than planned on a later wired segment if available. Also it would be sensible to avoid waste by retaining sufficient battery capacity to take all the regen from a planned long unwired descent, rather than having to burn off the energy in braking resistors if the battery is full. Advanced adaptive analytical methods might be used to develop, adjust and simulate charging plans and effects on the railway and wider supply networks (whew managed to avoid using the AI term!).

These discussions are 100% happening in the UK, not least because there's already a system explicitly designed for BEMU charging off the wires being delivered (Cardiff Valley Lines). I was involved in a project in 3-4 years ago that was talking about geofencing charging areas for BEMUs to reopen a freight line to passengers.

CVL is a very important project for demonstrating battery techniques. They already have some passenger running experience with the FLIRTs.

 

[edwin_m]

RSSB did some limited research into the current levels that could be transferred under project T1185 (Current Limit at Standstill). The testing demonstrated that a conductor beam with a 120 sqmm copper wire could allow as much as three times higher power transfer than standard contact wire over a 10 minute period without exceeding the temperature limits for the pantograph carbons or conductor wire.

Surprisingly a twin wire setup didn't achieve double but it still facilitates a higher power transfer.

Ultimately the rating of the transformer will determine how much power can be transferred.

https://www.rssb.co.uk/research-catalogue/CatalogueItem/T1185

The contact area would therefore be the same as with a normal contact wire, but the extra power is presumably because conductor bar just above it creates a heatsink and limits the temperature increase.

 

[zwk500]

RSSB did some limited research into the current levels that could be transferred under project T1185 (Current Limit at Standstill). The testing demonstrated that a conductor beam with a 120 sqmm copper wire could allow as much as three times higher power transfer than standard contact wire over a 10 minute period without exceeding the temperature limits for the pantograph carbons or conductor wire.

Surprisingly a twin wire setup didn't achieve double but it still facilitates a higher power transfer.

Ultimately the rating of the transformer will determine how much power can be transferred.

https://www.rssb.co.uk/research-catalogue/CatalogueItem/T1185

Is a bespoke setup above platforms more or less desireable and capable than a plug-in shore supply?

 

[N1]

Is a bespoke setup above platforms more or less desireable and capable than a plug-in shore supply?

A bespoke setup, most likely 25kV AC overhead, is more capable of providing recharges at branch terminuses with short turnarounds, and allows charging and electric running on already electrified areas.

Of course shore supplies were fine for battery trains in operation for almost all of the 20th century. Liquid cooled CCS chargers are currently limited by standard to 500kW per charger (500A at 1000V).

A 'bespoke' setup could imply cost, or it could be a couple of standard inverted pantograph chargers. The largest of these I have seen is 600kW, so would need a couple at each station. Obviously this layout limits you to a particular unit.

 

[HSTEd]

A bespoke setup, most likely 25kV AC overhead, is more capable of providing recharges at branch terminuses with short turnarounds, and allows charging and electric running on already electrified areas.

It's probably also going to be substantially more expensive than a charging station which will likely be a commodity item on account of goods vehicles of various types though.
I think trains willl probably need to have both.

Of course shore supplies were fine for battery trains in operation for almost all of the 20th century. Liquid cooled CCS chargers are currently limited by standard to 500kW per charger (500A at 1000V).

A 'bespoke' setup could imply cost, or it could be a couple of standard inverted pantograph chargers. The largest of these I have seen is 600kW, so would need a couple at each station. Obviously this layout limits you to a particular unit.

The megawatt charging system standard being developed for lorries will supposedly be capable of 3.75MW using liquid cooling.
As a moderate voltage DC spec it will also be much easier to bypass any transformer rating/weight issues.

 

[Nottingham59]

Is a bespoke setup above platforms more or less desireable and capable than a plug-in shore supply?

As far as I can see it's just a conductor bar like they have on Thamelink at St Pancras, but maybe a bit thicker. Hardly "bespoke". But I do agree that all new BEMUs should have plug-in MCS charging capability as well as overhead AC.

 

[edwin_m]

Is a bespoke setup above platforms more or less desireable and capable than a plug-in shore supply?

A shore supply would need someone to plug and unplug it, which eats into the charging time and creates various hazards whether it is at platform level or at track level. Bi-modes operating in 25kV areas will already have a pantograph, which can quickly be raised and lowered when needed.

 

[MarkyT]

Whatever solutions are chosen, they really need to ensure an immediate automatic connection for in service charging at platforms and turnback sidings so no valuable charging time is lost. Plug-in might be more suitable for overnight stabling locations where there are crew available who can plug and unplug units as part of preparation and disposal duties. Overnight stabling would not typically require very fast charging, but the total load for a large yard could be considerable.

 

[zwk500]

A shore supply would need someone to plug and unplug it, which eats into the charging time and creates various hazards whether it is at platform level or at track level. Bi-modes operating in 25kV areas will already have a pantograph, which can quickly be raised and lowered when needed.

The Plug-in stand would be fenced off at the platform end, hardly a major hazard, and would be 'operated' by the driver. Or, with a bit of effort, you could have a charging head under the conductor and couple to the buffers (obviously less practical in through platforms).

 

[Brubulus]

The Plug-in stand would be fenced off at the platform end, hardly a major hazard, and would be 'operated' by the driver. Or, with a bit of effort, you could have a charging head under the conductor and couple to the buffers (obviously less practical in through platforms).

The GWR solution seems much better for short turnarounds, though the simplest solution is a short section of 750V DC rail.

 

[MarkyT]

The GWR solution seems much better for short turnarounds, though the simplest solution is a short section of 750V DC rail.

A continuous rail or wire can cater for variable length trains and any number of shoes or pantographs. The Greenford solution with short conductor rails currently can't cope with a double unit. If two units ran coupled on any branch, say on the Henley branch at Regatta time, this could require both termini to have charging stations at the buffer stops so both units could take a charge alternately. Alternatively, a second length of rail might be provided in the appropriate position for the second unit, but that could limit the possible formations, and add extra switching complexity to prevent the second set of rails being energised if there was no second unit parked directly over it.