Published on

DFS: Return of the Flexing

Authors
  • avatar
    Name
    Archy de Berker
    Role
    Co-founder @ Axle
  • avatar
    Name
    Karl Bach
    Role
    Co-founder @ Axle

Last winter, the Demand Flexibility Service (DFS) delivered 3.7 GWh over 16 events. More than 2.6 million households and businesses participated, reducing peak demand and avoiding emergency use of coal power plants. At Axle, we contributed over 60 MWh of flex, marshalling thousands of batteries & EV chargers and paying >£120k to consumers in the process.

For 24/25, the DFS will evolve further. The Electricity System Operator (ESO) forecasts that this winter will be less strained than last year's, with a healthier margin of supply over demand. Nevertheless, the ESO remains committed to the DFS as a fantastic way to drive engagement with flexibility, and an important stepping stone to the 30 GW of flex they’d like to see by 2030. Here are the key changes the ESO is proposing for this year, and what they mean for distributed behind-the-meter assets.

From enhanced action to in-merit

DFS has historically been an enhanced action product for the ESO, which is grid-operator speak for “a big red button”. The aim was to provide a new tool which the ESO could use when the gap between supply and demand was worryingly slim (known as high loss of load probability). This special status allowed the ESO to pay higher prices for DFS compared to those paid for conventional grid balancing services, such as those provided by gas power plants or grid-scale batteries.

Since the outlook for winter 2024/5 is fairly rosy, they won’t be using DFS as an Enhanced Action product. Instead, it’ll become an in-merit margin service. In-merit means that DFS will be compared to the range of other options for procuring energy at times of tight margin. Intriguingly, the ESO have also laid the foundations to use DFS all year round: their proposal is not restricted to winter.

In-merit dispatch
In-merit dispatch
  1. Offers are sorted by price
  2. The required volume sets the acceptance price
  3. Volumes below that price are accepted, above are rejected

Is this good or bad news?

This will probably mean lower prices - certainly lower than the £3000/MWh we saw for test events in 23/24. However, it will also offer more opportunities for dispatch, since DFS participants will be called upon whenever they can provide power at a competitive price.

For those who are manually instructing flexibility, this will probably be a net negative: more frequent, less valuable dispatches will increase operational costs and potentially disenfranchise users. For assets like batteries and EV chargers which can provide automated flex, this change will likely increase the total value for minimal additional effort.

Other changes

There are several changes which fall out of the transition from an Enhanced Action to an In-Merit Margin Service. These which will bring DFS in line with the other services used to balance supply and demand:

Stacking will be allowed

Last year, DFS wasn’t stackable with other flexibility services: if you were participating in DFS, you couldn’t do anything else with the asset. Happily, that will change for this year, allowing assets to participate in the Capacity and DNO Flexibility alongside DFS.

This is clearly a win. There’s also some detail on how participation in other markets will be subtracted out of the baseline calculation, so you don’t need to worry about cannibalising value from DFS by participating on load-shifting programs.

Asset metering opened up

Last year, most participants used boundary meters (i.e. smart meters) to quantify their participation. Asset metering - submitting measurements from an EV charager or battery - was technically allowed, but only in cases where the boundary meter was Half Hourly Settled1. Since only 20% of consumers are currently Half Hourly Settled, this effectively ruled out this approach.

This added friction for non-suppliers, who had to jump through hoops to get access to boundary meters on behalf of customers. Stripping away this complexity is a major win for companies with access to asset data - such as hardware manufacturers - who can now participate with the data from devices with no extra data flows.

No guaranteed test events

In 2023/24, we saw 12 test events, of which 7 were competitively priced. At this point, the ESO is satisfied that DFS is a robust and competitively priced product, so they're not planning any test events in 2024/5.

DFS has graduated from the minor leagues, and it’s ready for the big time.

We’ll only see within-day dispatch

Dispatches last year ranged from day-ahead to within-day with varying degrees of notice. For 24/25, DFS will only be procured within-day, in line with other balancing products.

Performance penalties

Payment for DFS in 23/24 was proportional to delivery: there were no explicit penalties for underdelivery.

As the ESO looks to reshape DFS from a maverick knight-in-shining armour to a reliable footsoldier in the battle to balance the grid, it’ll be tightening up the rules on delivery. They propose the following:

  • full price if you deliver between 50% and 120% of the procured volume
  • reduced price if you’re between 25 and 50% of procured
  • no payment if you deliver <25%

This was inevitable; if we want demand-side flex to be a critical part of grid balancing, we need increased certainty about how much power we’re actually delivering.

This could be challenging for participants eliciting manual turndown from households & businesses. Voluntary turndown is a complex function of notification strategy, time of day, and what is going on in people’s lives (exhibit A: the kettle spikes from the Euros last week). In our experience, discharge volumes from API-controlled assets is highly predictable, and the increased premium on controllability stands to benefit participants with more reliable dispatch.

What happens next?

DFS 2024/25 Consultation process

Over the next few months, the ESO is consulting on these proposals, with the intention of submitting to Ofgem in September. Assuming Ofgem approves, we’ll see these updates in force from November 2024.

Participating in DFS this winter

Axle Energy was the largest automated participant in DFS 2023/4, helping clients like GivEnergy deliver > £120'000 of rewards to customers. If you have distributed assets like batteries, heat pumps, or EV chargers, we’d love to chat about how we can help you participate in DFS 2024/5.

Footnote 1: Half Hourly Settlement

Most UK customers are not settled on the basis of smart meter data. Instead, suppliers use standardized daily profiles and multiply through by annual consumption to obtain 30 minute estimates. This is changing with the introduction of Market Wide Half Hourly Settlement, which should be complete by December 2026.