Rail Engineer has written extensively about the development of Battery Electric Multiple Units (BEMU), and Discontinuous Electrification. In Issue 215 (Jul-Aug 2025), our article about the Cardiff Valleys explained the importance of system design to ensure that BEMUs can get around the discontinuous OLE without running out of power. Another article which covered an IMechE event quoted numerous values for battery capacity and projected range.
More recent Permanent Way Institution (PWI) and Railway Industry Association (RIA) events have suggested approaches which might lead to a level of standardisation. But questions remain: is discontinuous electrification cheaper than continuous electrification when considered in whole life, whole system terms and, if so, where?
What follows are your writer’s views about many of the issues to be resolved based on extensive experience as a systems engineer. It takes into account the thoughts of electrification expert Garry Keenor as presented at a recent PWI lecture and proposals from Siemens Mobility presented at a recent RIA event and, indeed, RIA’s views.
All commentators have expressed the view that heavily used main lines and those carrying extensive freight should receive continuous electrification. This is simply because battery bi-mode locomotives do not have the range or speed that would be needed. That said, a battery capability to allow operation over short discontinuities, really challenging obstructions, or last mile situations such as freight yards and depots might be sensible – see the rolling stock comments below.
As already stated, any form of discontinuous electrification requires a high order of system design, using an operationally led specification including rolling stock, electrical supply, electrical distribution, power changeover, civil structures, and signalling issues together with the geography of the lines concerned such as frequency of, and distance between, station stops as well as gradients.
Taking the various issues in turn:
The operator
The concept of operations should be at the front and centre of modelling and design. For example: what is the service pattern and how might it change in the reasonably foreseeable future? Where and for what duration are the dwell times for static charging? How much contingency should be built into the system: e.g., unavailability of a static charger for a day or virtual elimination of layover time due to disruption? Or, what if a line blockage means the train has to turn back just before its next charging opportunity?
Drivers will need to know whether a journey can be undertaken or completed before setting off, requiring some sort of connected driver advisory system. Such a system could determine route conditions and battery state and give a go/no-go to the driver.
Passenger rolling stock
Although there may be exceptions at the margins, it is generally true that adding traction batteries to an electric train produces a less satisfactory product. BEMUs are:
- Less energy efficient than straight electric.
- Heavier and therefore increase attrition on the track system shortening its life.
- More technically complex.
- At higher risk of a serious fire.
- More costly in whole life terms.
Another factor (and a feature of the current Northern Trains procurement) might be a short-to-medium term requirement to provide diesel-electric bi-mode trains with the potential to be converted to battery electric in the future. In your writer’s experience, such conversions always cost more and present more challenges than can be foreseen. These might include standards moving on and assumptions about conversion made during design that are no longer valid when the time for conversion comes. Some train types such as the Stadler bi/tri-mode FLIRT design, where the diesel engines and batteries are in a separate pod between passenger coaches, might offer easier conversion but come with the disadvantage of a longer train for a given passenger capacity.
There can also be restrictions on cascade unless there is some standardisation on battery capacity. It’s no good cascading a fleet with a 30-mile battery range onto a route requiring a 50-mile range, for example. Siemens Mobility has suggested a standard 110mph train with a reliable range of 50 miles at the end of a 20-year battery life. Might these parameters be a good base around which to design the appropriate mix of continuous, discontinuous electrification and charging stations? Perhaps the normal requirement could be a battery capacity of XX kWh per tonne of train mass with standardised charging arrangements.
In passing, it is worth mentioning that there is a trend of adding batteries to new trains running on continuously electrified lines. New trains for both London Underground and Nexus (Tyne and Wear Metro) have batteries to allow short duration movement if the traction supply is lost and for movements around the depot. Southeast Trains’ (SE Trains) Networker replacement stock is proposed to have batteries too with a range of approximately 15km, Rail Engineer understands. Batteries on EMUs could also help deliver economies and improve safety if OLE or third rail could be removed from depots – except in dedicated charging locations. Again, Cardiff Valleys leads the way with very simple OLE conductors for charging the batteries of Class 398 trains in Taff’s Well depot whilst avoiding the usual knitting of the wiring in the depot’s complex points and crossings.
Electricity supply
Whether powered directly from the OLE/third rail or battery, BEMUs need an external power supply for at least part of the journey (moving or stationary), and the intermittent external supply must have the aggregate capacity for the whole journey. Providing this power from the very high voltage system (e.g., 400kV, 275kV) can be a challenge. The nearest connection point might be a long way away and, in some places, there might not be any connection point available. Even if there is, it could take up to a decade to obtain the connection and it is always costly. Even today, this type of constraint means that there are restrictions on the number and size of trains that can draw power on the East Coast Main Line in Northumberland, and that was electrified 40 years ago. Indeed, more bi-mode trains operate with diesel power north of Newcastle in the enhanced December 2025 timetable.
Another technique is using so called Static Frequency Converters (SFC) which take the three-phase supply and convert it to single phase 25kV (SFC is something of a misnomer in the UK as the frequency is not converted). SFCs provide a more balanced load for the grid than the conventional supply approach which means they can be connected to more numerous 132 kV supply points.
An alternative, suggested by Siemens Mobility, involves connecting to, for example, the 11kV or 33kV three phase supplies and converting this into single phase 25kV using a device which Siemens Mobility calls a Rail Charging Converter (RCC) and which could deliver up to 2.5MW. Whereas very high voltage connection points are rare, 11kV and 33kV supplies are very frequent. As the name implies, these are aimed at static charging locations for relatively low powered trains – e.g., branch lines. The ability of these connection points to support the RCCs traction power demand is another issue to be resolved.
Static charging
Battery charging stations might be located at terminus or bay platforms allowing a train to be charged between its inbound and outbound journeys. Challenges include:
- Layover time.
- Battery charging time.
- Maximum current draw through the OLE whilst stationary.
- Operational factors such as arrangements should a train be late on its inbound journey.
Rate of battery charge is also a factor, as very high charge rates to enable a short charging time can reduce battery life. Equally, high charge rates can stress the power supply which might dictate a more complex charging station using static batteries trickle charged from the grid supply to provide a high charging station-to-train charge rate (demonstrated in the GWR trial at West Ealing).
Despite all precautions, the grid supply available at the terminus location might not have the required capability necessitating a design revision. Moreover, a static charger is as much a part of the traction system as is a feeder station on a continuously electrified line and needs to be engineered to at least the same level of reliability and dependability.
Clearly, if a reasonably common rolling stock fleet is the aim, then a common means of charging the batteries is required. Two systems have emerged in the UK so far: (i) the Vivarail/GWR fast charging via collector shoes/conductor rails in the four foot; or (ii) via 25kV overhead line in Cardiff. Separately, charging via 750V DC has been implemented in Liverpool.
Modelling will show where trains need to be continuously powered while charging during some parts of the journey supported by static charging at, say, termini and depots. Once again we argue that the infrastructure which inevitably varies between different routes should be designed around a relatively standard train. Hence, there should probably be just one or a maximum two designs of connection between trains and electrical supply: 750V DC third rail and 25kV AC in their respective areas.
Where does this leave the VivaRail/GWR system? It might be a casualty of innovation despite its success in raising the profile of independently powered electric trains, although its principles could be applied to conventional third rail charging stations or islands.
In-motion charging
Sections of 25kV OLE need to have enough capacity to both power the train and charge the batteries at the same time. If the train uses existing OLE for this purpose on part of its journey, the OLE and power supply might need to be upgraded to cope with the additional battery charging load. These sections also need to be long enough to deliver enough charge to allow the train to safely reach the next continuous power section or static charger.
OLE and obstructions
It is generally accepted that erecting simple plain line OLE is not the most expensive part of electrification. It is the features that interrupt plain OLE (e.g., junctions) and obstructions that require reconstruction (e.g., bridges, tunnels and platform canopies), that add significant cost.
Recent work by Network Rail has produced techniques that allow very small electrical clearances. This work has, for example, led to a markedly reduced the number of bridges needing attention on the Midland Main Line electrification. If the catenary does have to be interrupted for an obstruction such as a tunnel, there must still be a means to provide the 25kV supply either side of the obstruction. This might mean providing a 25kV buried cable which is comparable in cost to providing OLE, albeit saving the cost of tunnel works. However, in an intermittent electrification scheme, a tunnel with sufficient clearance for OLE might present a way of implementing a charging island, reducing the cost as there would be a reduced need for foundations and masts to support the OLE.
Power changeover
Ensuring that the pantograph is lowered before the end of a catenary section is generally automated, but if all else fails, the final control is in the hands of the driver. Therefore, power changeover must not occur at locations where the driver already has a high workload, for example, when approaching signals. As an example, ideally the train would change from OLE to battery before leaving the main line onto a branch line, whereas the detailed study might conclude that it is better to change over after the train has passed over the junction.
In other places, pantographs might have to be lowered much further from a catenary free obstruction than has been assumed in the concept design. Additionally, it will generally be important to return to OLE power as soon as possible to avoid unnecessarily depleting battery range. Changeover catenary sections tend to be bespoke designs and, if speeds are high, a considerable length of track might be needed. Standardisation will be important to avoid captive train fleets.
Signalling
Discontinuous electrification does not remove the need to immunise signalling and telecoms systems, which will still be required around charging stations and electrified islands, and potentially elsewhere. There may also be signal sighting issues to resolve.
Modelling
Although last in this description, modelling is an integral part of the design and must be started early. As modelling assumptions are hardened into requirements (for example, battery range in the train specification) any further modelling will not be able to vary those parameters.
Taking all the factors described into account, modelling can help determine and optimise the various parameters of all components of this tightly coupled system. Even a standard approach to train battery capacity will have to be modelled route by route: a flat route with a long distance between stations, for example, will deliver a longer range than a hilly route with frequent stops. Modelling needs to take account of reasonably foreseeable changes to the timetable too, then a proposed charging station/OLE layout can be designed as a concept, taking account of the obstructions mentioned above and the modelling re-run to assess whether the train’s batteries continue to remain in a good state of charge.
Inevitably the layout will need refinement before an optimised system solution is achieved. As design detail progresses, further changes to the infrastructure design might need to be modelled before adoption as, by then, the train design will be fixed.
Standards and safety
Many recent schemes have attracted significant cost for compliance demonstration and safety assurance. The cost will be higher where safety has to be demonstrated through ‘explicit risk estimation’. This was the experience in Cardiff. Thus, it is important to create standards or assured standardised design templates that can be applied widely for all sub-systems.
So where is discontinuous electrification the right answer?
What has been written so far might sound discouraging. This writer believes that the whole life cost of discontinuous electrification could be higher than a continuously electrified line. There are, however, exceptions in these cases (credit to AtkinsRéalis’ Garry Keenor):
- A single discontinuity to deal with a problem with no practical engineering solution.
- Last-mile branch line capability in co-ordination with continuous electrification on the main line.
- Multiple discontinuities on a mixed-use passenger and freight railway as part of a longer-term strategy to close the gaps for freight (some form of bi-mode freight locomotive in the interim).
- Short, self-contained lines with no freight where island charging can be provided.
- Multiple discontinuities on a self-contained passenger-only network, albeit this is unlikely to pass a whole life cost test.
We have seen proposals for discontinuous electrification on East-West Rail where the current sole user is freight, so this only makes sense if there is a plan to eventually close the gap (Point 3). In Scotland, a recent announcement suggests an integrated plan to purchase BEMUs and provide discontinuous electrification on lines where freight is unlikely (Points 2, 4, and 5).
Paying for everything
This remains the key challenge. The current stated position on further electrification is stark: “there is no money”. Yet several train operators have invited tenders for hundreds of new vehicles, mostly bi-mode units in one form or another. These vehicles will be lease financed, so the cost will be spread over many years. Rail Engineer wonders whether electrification could be financed the same way. After all, we have seen depot extensions including fuel filling systems for diesel trains financed by ROSCOs, so perhaps the electrical equivalent of a diesel’s fuel filling systems for (B)EMUs (i.e., charging stations and/or OLE) could be financed this way, irrespective that some of the ‘charging systems’ might be many miles long!
That said, the business cases so far presented for discontinuous electrification are based solely on capital investment costs which show only marginal savings compared with continuous electrification, greatly sensitive to the amount of OLE. Once the OLE sections get too much over 50%, continuous electrification wins. In whole life cost terms, continuous electrification wins emphatically.
As for freight, discontinuous electrification is nonsensical. Batteries will only ever be practical for last mile operation and not for main line freight movements. Diesel bi-modes have relatively weak engines. Even the mighty new Class 99 with a fully laden train will, on electric power, be able to romp up steep gradients at more or less maximum freight speed whereas on diesel it will be lucky to be running at 25mph by the summit. That is, of course, a real problem on a mixed traffic route.
Strategic view required
Discontinuous electrification can take many forms. It is important to choose the right approach for each route. With the emerging proposal for regionally integrated organisations in Great British Railways, ideally someone centrally will take a strategic approach which is operationally led and enables the application of templated designs.
This applies especially to rolling stock to avoid the risk of trains being locked to a route for their whole life irrespective of need. A strategic system view is also required for financing it all rather than treating each part of the system in isolation.
Why standardised bi-mode trains?
In 2023, Rail Engineer wrote about the challenges faced by East Midlands Railway when integrating Class 170 trains from three other operators into its own fleet. Summarising, the most significant technical issues were that some units had different autocouplers; customer information scripts had to be changed; and there were differences in some control panel layouts. Whilst significant, these were comparatively straightforward to change. Apart from that, all the basic mechanical features were the same.
Imagine a situation in 2050, when it is decided to transfer 15-year-old ‘standard’ BEMUs built by (fictional supplier) UK United Trains from three routes to a fourth that already has some of these trains. Imagine the route engineering manager’s consternation when they find that they have four types of trains, three of which have less battery capacity than their own fleet which is insufficient for their routes and, moreover, they have different charging requirements. Clearly the trains and chargers could be modified, but the cost would be high.
Rail Engineer thanks Gary Keenor, Peter Dearman, and David Clark for their input.
Image credit: