This project will assist with the facilitation of rapid EV charging equipment by providing appropriate network connections where they are needed, whilst making optimal use of the available network capacity.
The connection costs for rapid EV charging facilities are a critical consideration for developers. Provision of rapid charging facilities is key to the uptake of electric vehicles (particularly for fleets and urban delivery services) and hence high connection costs in some areas could be seen as a potential barrier to the uptake of EVs. Users must be able to charge their EVs in a manner that is at least as convenient as current refuelling methods, which means minimum dwell time, or opportunity charging where vehicles may stop.
DC share is a smart DC network solution that facilitates rapid charging in constrained areas by using the available latent capacity across a number of substations. The solution will mesh a number of distribution substations, with DC converters and high capacity DC links. The DC system will then determine the best point to draw power from for the charger and it will also support heavily loaded transformers.
Benefits
There are significant benefits to implementing power systems with more flexibility, enabling dynamic optimisation of use of existing assets, and providing capacity where it is needed. Once implemented, the DC Share solution will result in:
- Significant financial, capacity, and carbon benefits associated with the EV charging network and optimisation over multiple substations – These financial, capacity, and carbon benefits have been quantified through business case modelling, as described below:
- £162m in direct financial benefits up to 2050 across GB, which will benefit the customer though network savings resulting in lower customer bills, and through enabling access to more infrastructure at a lower cost.
- 1,800 MVA capacity released up to 2050 across GB
- 26,000 tCO2e of direct savings up to 2050 across GB
- Substantial indirect carbon and environmental benefits through supporting the uptake of EVs and the connection of LCTs – The availability of charging is a significant enabler of the uptake of EVs, and the release of capacity through real-time active management can enable the connection of LCTs at a lower cost. There are substantial environmental benefits, including 21 tCO2e carbon savings through electrification of each passenger car and 67 tCO2e through electrification of each taxi (based on average mileage and fuel/energy consumption figures). This reduction in carbon emissions will have substantial impact on air quality in the GB as the vehicle part transfers, and DC share will facilitate this through provision of rapid charging facilities.
- Minimise the impact of significant clusters of rapid charge points on the network - As part of the DC Share solution, rapid charging points can be managed and optimised as part of the wider DC solution, taking advantage of flexibility in charging time and minimising the peak demand on the network.
- Increased network flexibility – The provision of flexible methods enables increasing uncertainty to be managed more effectively, optimising power flow in real time to react to changing network demands and providing real-time controllable support to the wider AC network.
- Future-proofing of the network infrastructure and avoidance of stranded assets – The solution can provide future-proofing through enabling the connection of future DC loads, generation and storage to established DC Share solutions. The topology of a DC Share solution can expand and adapt over time to meet the changing future needs of the customers. Under-utilised substations can be connected to the DC network, allowing them to provide additional capacity to the system and enabling avoidance of stranded assets.
- Network control benefits – Additional network control benefits using Power Electronics may be enabled through the solution, such as the ability to actively manage network voltage and power flows, which can offer customers improved quality of supply which can be maintained through changing network conditions.
- Reduced Losses – Charger losses are lower in DC Fed EV Charge points, by virtue of the simplified architecture of the devices. System losses are also lower in the DC Share case than in the “Soft Open Point” and “Solid State Transformer” cases, due to the reduction in conversions between AC and DC. Losses can also be minimised on the DC Share system, by actively managing power flows (when possible) to minimise current in the parts of the network most susceptible to losses, and by switching off converters when not required for use.
DC Share is well aligned to the Western Power Distribution, Electricity North West and overarching ENA innovation strategy, of facilitating change. DC Share will provide a coordinated approach to the deployment of EV charging points, whilst also using smart technologies to maximise capacity within the power system.
Learnings
Outcomes
Our learnings about LVDC equalisation from this project can be distilled into the following insights about their application.
Enablers prior to the decision to build an equalisation network
We showed that the amount of capacity that is available in an LVDC equalisation network is dependent on the demographic of the customers, but also that it is essential to know the detail of the daily load cycles before understanding how much capacity is available across a selection of substations.
We also showed that use of a simple model using typical MPAN settlement profiles is an unreliable method to estimate the capacity available in an equalisation network as it tends to underestimate the capacity available for sharing.
This means that to be able to deliver LVDC equalisation networks at scale, there would be a dependency to have measured the daily load profiles of candidate substations before deciding which substations shall be used for the equalisation network. Because estimation of load profiles based on settlement profiles is unreliable, it means that LVAC modelling needs to have been in place at candidate substations for a suitable period of time to fingerprint the capacity available for sharing.
This in effect means that an enabler for LVDC equalisation networks is to have LVAC substation monitoring in place to be able to decide whether an LVDC equalisation network would make sense. NGED’s position on LVAC monitoring is that we believe it adds value at sites that are at high risk of having their ratings exceeded, but installing monitoring at sites that are likely to remain underutilised is a poor use of customer money.
In addition to the need for visibility of LVAC load profiles, we believe that DNO’s would need to work with suppliers to establish a new capability to be able to model load flows and fault levels within potential LVDC equalisation networks.
The need for the right substation fabric, in the right location with the right type of capacity
The work done to date demonstrates that delivery of an LVDC equalisation network requires several conflicting factors to be resolved.
- The LVDC equalisation network needs to be proximate to where developers would like to construct EVCPs. Outside of trials, this is not something we would typically expect the DNO to have direction influence over
- We demonstrated the practical problems of installing the DC cable route across a town centre. In particular, it was observed that substations that look promising from a fabric perspective, may have a challenging cable route
- We demonstrated that the protection philosophy will place a limit on the total amount of cable that can be allowed in an LVDC equalisation network. This is in tension with the cable route logistics needing to circumvent engineering obstacles by going around them
- We demonstrated that less than one in in two substations will be able to host the infrastructure required to convert AC to DC electricity. This will mean that the LVDC equalisation network will need a bigger footprint to connect the substations that can host GTIs. This is in tension with the protection philosophy need to restrict maximum circuit length to be able to link up with substations that have sufficient space or assets.
- We demonstrated that not all substations will have the right load profile for equalisation. This will be in tension with the protection philosophy’s need to limit cable route length.
We believe that these tensions make the prospect of LVDC equalisation networks unsuited to mass application.
Efficient network development
The minimum commitment that can be made towards an LVDC equalisation network would be installation of two GTIs at two substations, one supervisory controller and the DC cabling, in addition to any EVCP. This would be required, even if there is only one overloaded substation or EVCP site requiring connection.
We also believe that the maximum allowable footprint of a LVDC equalisation network would be in the region of a maximum cable route of 1227 metres, after which a new and separate LVDC equalisation network would need to be established to avoid insufficient protection coverage.
These limitations do not leave much flexibility for development of the equalisation network in between the two extremes.
As also shown, LVDC equalisation networks will require EVCP customers to accept a certain level of curtailment. Any load growth or new connections on the AC side of the infeeding substations will increase the amount of curtailment experienced by EVCP over time. In the event that the curtailment experienced becomes untenable, the only strategy that the DNO will have will be to either reinforce the existing HV/LV transformers or alternatively establish new HV/LV transformer substations to relieve the existing substations.
We showed that the BaU solutions which establish new HV/LV transformers in first place are cheaper upfront. As a result, we consider that selecting an LVDC equalisation network as the first step in the network development path leads to the greatest potential investment regret, rather than the minimising the investment regret.
Customer Proposition
We have shown that use of LV equalisation networks and DC connected EVCP does no not necessarily provide an equivalent level of service to BaU connections. In particular:
1. LVDC equalisation networks are likely to provide a slower connection than comparable BaU connections with a greater delivery risk
2. LVDC equalisation networks are likely to provide a more expensive overall connection scheme than an equivalent BaU connection
3. LVDC equalisation networks will not have the same level of redundancy as a BaU connection with equivalent capacity
4. LVDC equalisation will require the capability to curtail customers.
It should be remembered the DC Share project was commissioned prior to the output of the Significant Code Review (SCR) regarding system access and charging. This means that the customer barriers that DC Share sought to respond to have moved significantly. Firstly, in the FSP it was assumed that EV charge point developers would have to fund any reinforcement but as a result of the SCR, all reinforcement costs for EV demand will be socialised. This means that the upfront cost of reinforcement expected by EVCP developers will be approximate to the connection assets. The FSP also assumed that it would be acceptable for the DNO to curtail customers if they didn’t wish to fund reinforcement. As a result of the SCR, DNO’s will now be required to set a limit to the amount of curtailment a customer will experience. Both of these changes substantially alter the aims of the original DC Share FSP. This is because reinforcement cost socialisation mean that traditional reinforcement approaches are no longer the barrier to decarbonisation that they were.
The requirement for DNO’s to limit the curtailment experience would make the use of LVDC equalisation networks problematic because there would need to be a development strategy to reduce curtailment. We showed that the protection system will limit expansion of the equalisation network meaning it is unlikely additional DC cable route and GTI’s can be added. Therefore, the only strategy the DNO would have to reduce curtailment would be to create new HV/LV AC substations and construct new AC cable routes to transfer the incumbent AC load across to these new substations.
Lessons Learnt
Network concept development
This project demonstrated the value of ensuring that there is a clear grasp of the overall network concept and who the system is expected to help. This concept needs to include the following:
- The required customer experience, including business case, connection experience, ongoing capacity availability and service levels
- Expected loading on assets and the operational events which mark the limiting conditions
- The protection philosophy confirming whether acceptable levels of protection sensitivity and clearance speed will be delivered and under which conditions the system reaches performance limits
- The expected device performance and the boundaries which mark the limiting conditions
- The means to ensure that acceptable performance of the earthing system will be delivered
- The means with which a safe system of work can be embedded into the system operations
- The likely constraints to be expected at the installation sites
To some extent, this learning has already been adopted into the future Strategic Innovation Fund (SIF) structure and the DC Share project would have benefitted from this structure. For example, the discovery phase could have reviewed approaches to supplying EVCP and the viability of LVDC equalisation networks against alternative technology. The alpha phase could then have been used to develop and verify the network concept for LVDC equalisation networks then progressing through the beta and trial phases, had the concept still appeared viable.
When developing new concepts for LV networks, innovators must consider the diversity across substations that will be experienced. For this reason, network concepts should be stress tested to understand in how many instances they are expected to be viable instead of depending on a small number of case studies.
Reducing construction delivery risk
The DC Share project infrastructure was to be located in the middle of a town centre. Construction project in urban environments do often come with delivery constraints that a network operator has little influence over (limitations on road opening periods or temporary one way systems etc). These delivery constraints increase the cost and timescale of delivering the trial. When planning future innovation projects consideration should be given as to whether there are cheaper methods or locations to gather the same learning.
Adoption of trial assets into daily operations
The project intention was for the LVDC equalisation network and LVCP to be adopted into daily network and car charging operations. Had the network performance or the EVCP performance begun to deliver unacceptable performance beyond the warranty period, there was a strong likelihood of stranded assets.
In the future we recommend that any application for SIF funding where new assets are to be trialled should contain an indicative business plan and risk management plan for the assets for ongoing operations after the trial. These documents should demonstrate how the post-trial operations of the new infrastructure are expected to technically and commercially protect customers. In the event that there is too much uncertainty as to whether post-trial operations would be successful, then the trial methodology should be reviewed to minimise any regret if the trial demonstrates a non-viable proposition.
Updated view of base case
Since 2019, the volume of connections for clusters of rapid chargers have grown from few instances, to become regular part of the connections sector. This has allowed us to update the BaU costs beyond the FSP assumption.
We observe that it is common for new connections to rapid charging clusters to be delivered by an independent connections provider. We also observe that there is a preference to have a HV point of connection, with the EV clusters fed from a dedicated HV/LV substation located on land provided by the EV cluster developer adjacent to the chargers. We believe that this is partly due to ease of construction, but mostly because this form of scheme design has grown to be the least cost connection for the customers’ needs in comparison to an LV point of connection. We believe that connection projects of this nature can be delivered within 3 to 9 months from offer.
Learning regarding the method case
Since embarking on the design of the DC Share project, we have learnt more about the cost base requirements to establish a DC equalisation network, including:
1. In 60% of instances, the LV point of connection will require remodelling, at a cost of between £6k and £25k not including the cost of any cooling or land purchases that would be required.
2. It would be prudent to allow for a proportion of the cable route to be installed in the public carriageway.
3. To have an equivalent resilience of supply to a BaU connection, the infeed’s into the equalisation network need additional redundancy.
Learning regarding alternative innovations
Since the commencement of the DC Share project, two NIC projects fully delivered their output. These two projects were NGED’s OpenLV project and ENWL’s Celsius project. Both of these projects investigated the use of site-specific thermal ratings on HV/LV transformers. Both of these projects demonstrated that this technique could release significant additional capacity in HV to LV transformers. As an example, the Celsius project indicated that in a population of transformers:
- 40% of the HV/LV transformers could carry greater than 130% of nominal rating
- 38% of HV/LV transformers sites could have a rating enhancement of between 110% and 120% of nominal rating. The OpenLV project projected similar learning.
This learning is significant for two reasons. Firstly, it indicates an alternative path to avoid reinforcement of the HV/LV transformers (which was one of the value streams that DC Share intended to respond to). Secondly, to deliver site specific ratings, exactly the same kind of LV monitors would be required to be fitted to the substations, that would be required to device it sites were able to participate in a DC equalisation network.
To deliver the capability, the host DNO would be required to install a monitoring device in the substation with a suitable thermal ratings app loaded within it. These devices are laptop computer size and can be retrofitted into existing substations. Experience shows that NGED can install these in ground-mounted substations with few exceptions. We have already committed to install significant numbers of these monitors into our substations across RIIO-ED2.
The lack of engineering complexity to deliver this capacity uplift should be contrasted against the DC Share learnings about delivering LVDC equalisation networks. This exercise presented learning that there are different approaches to release HV/LV transformer capacity to customers.
It should also be understood from the DC Share learnings that before an LVDC equalisation network could be constructed, a monitoring campaign would have to be undertaken using the same technology that could deliver site-specific thermal ratings. This means that HV/LV transformer capacity uplift could be delivered without having to commit to the engineering tasks involved in the DC Share project. This approach could also be delivered without the complexity of combining different load profiles.
The cost of procuring the infrastructure required for this approach to increasing HV/LV capacity would be less than £5k per installation and would provide additional benefits in addition to the capacity uplift. We believe that this is relevant learning to the DC Share project as it shows how significant capacity uplift on HV/LV transformers can be released, without having to resort the expense of creating an LVDC equalisation network or installing GTI’s in substations. This technology can also be easily installed into existing HV/LV substations because of their small size. This observation illustrates the importance of ensuring that LVDC equalisation networks is contrasted against the results from other NIC projects that have recently published their final learning.
