Opening up balancing to batteries

Following the launch several months ago of the Electricity System Operator’s new Open Balancing Platform, which will eventually replace the Balancing Mechanism, Utility Week speaks to battery operators to gauge its performance so far. Although acceptances have increased and skip rates fallen, they say there is still some way to go until batteries are competing on even terms with large fossil fuel power stations.  

Issuing an instruction to a generator to raise or lower their output typically takes an engineer in the Electricity System Operator’s (ESO) control room around 40 seconds per unit, says Waqquas Bukhsh, a lecturer from the University of Strathclyde.

Before this engineers must decide which units to dispatch, selecting the cheapest available to meet the needs of the system, whilst also grappling with network constraints and the technical capabilities of the various types of assets.

This task was manageable when the Balancing Mechanism was the domain of large generators that were relatively few in number. But the opening of the gate to smaller assets and batteries, much more numerous for the equivalent size, means this is quickly becoming too much for humans to handle by themselves. Simply hiring more people would not solve the problem, says Bukhsh: “They needed to automate that process.”

The ESO’s solution is the Open Balancing Platform (OBP), which launched in December and will ultimately replace the Balancing Mechanism.

It’s hoped this platform, comprising both hardware and software, will help to address complaints by operators of smaller assets and batteries that they were being skipped over in favour of large fossil fuel power stations, despite being further down the merit order.

Ben Guest, manager of the Gresham House Energy Storage Fund, says the results so far suggest this problem has not yet been resolved. He prays the ESO are just on the cusp of an improvements and his comments “look a bit silly” in hindsight: “I’d prefer that than to be right in my caution.” But for now, Guest has “lost faith” that this will happen soon.

Bulk dispatch

At the core of the platform is a “bulk dispatch optimiser” written by Bukhsh whilst on secondment to the ESO. His work on the project originally started off as a research project funded by the UK Research and Innovation in 2021.

Bukhsh says the software acts as the “brain” of the platform: “A control room engineer describes a requirement that they need 100MW in 15 minutes. That goes into the bulk dispatch optimiser and what falls out is dispatch instructions.”

“All the BM rules are coded within the optimiser,” he explains. The programme takes technical information about the available units – ramp rates, minimum zero times, synchronous or non-synchronous – as well as commercial information – their bids and offers and the ESO’s demand curve.

It then stacks the units from cheapest to most expensive to produce a merit order and a set of dispatch instructions for the engineer’s approval. They can then either execute these instructions or “pick out some of the units and re-run the programme.”

“It releases control engineers’ time,” says Bukhsh. “They just need to sense check that everything looks okay. Just press a button and off go the instructions.”

For the time being, the optimiser is only being applied to two “zones” within the Balancing Mechanism – one for batteries and one for smaller units: “The reason it’s not applied to all units is because the automation needs are in these two zones where there are really large number of units with less megawatts”.

However, Bukhsh says the bulk dispatch optimiser is “agnostic to the unit type” and will eventually be applied to all units as the OBP replaces the Balancing Mechanism.

Transparency and skip rates

Prior to the launch of the platform, operators of small assets and batteries had complained that the lack of transparency over why decisions had been made had left them wondering whether they had been treated unfairly by engineers who were more comfortable dealing with large conventional power stations.

Bukhsk says the bulk dispatch optimiser is “an answer to people who complain about skip rates. We have taken out that human element. Now the computer is making a decision based on the costs which go in.”

He says the optimiser will also be published so it is “open source” and “people can implement it themselves and validate the results… They can do ‘what if’ analysis with bidding strategies.”

Yet, Guest is still not happy with the situation. He says batteries and other small assets do appear have been used more in order of merit but overall volumes have not improved enough: “Instead of getting the odd trade as we did before, we’re now getting loads and loads of trades but they’re absolutely tiny, with the overall volume still remaining disappointingly low.

“If the ESO turn around and say: ‘Look, it’s better than it was.’ Fine. But there are so many batteries in the BM waiting to be used. We’re all fighting over very little.”

He believes this may be because of the absence of smaller assets from reserve services such as Regulating Reserve, which the ESO uses to correct energy imbalances.

The ESO creates the reserve by accepting offers and bids from generators to create headroom and footroom. Guests says the reserving of these plants locks them into engineers’ “field of vision” and gives them confidence the assets will be available when called up. He says the mindset of control room engineers seems to be “if it’s not reserved, it’s not available.”

Guest hopes that the launch of the ESO’s new Balancing Reserve service, which will procure Regulating Reserve through day-ahead actions may help to solve the issue: “The hope for us is that the reserving environment shows up, which offers us an availability payment, which then means that significant volumes are put through batteries as a result of them being dialled up as you go into an evening peak for example”.

But he says none of the milestones the ESO has achieved so far has brought significant improvement in battery revenues. He said batteries are being commissioned “at a much faster clip than the tiny improvements that are showing up in the control room, if there are any at all.”

He continues: “We’re not asking for a specific number of trades. We’re not asking to make a certain amount of money. We’re asking to be used properly and in order of merit.”

“That’s my only request and I think it’s a completely reasonable one because I can see that they’re working hard at this, but so far, it’s not leading to the improvements that we need to see. And the lack of those improvements is causing significant concern in our marketplace.”

Guest says there is still no clarity over “why a battery won’t win despite being competitive against a gas asset. I just don’t know whether there’s a situation that causes batteries not to be used for reasons that are yet to be explained.

“There’s also the possibility that National Grid takes an ultra-cautious approach for reasons I cannot understand either, given how ready we are and how capable the batteries are.”

Teething issues

Paul Mason, chief investment officer for battery developer Harmony Energy, says there were some “teething issues” when the OBP was launched: “It was live for a couple of days and then it was paused for a while.”

He says there were several occasions where the algorithm accepted some “extremely high” offers from batteries: “I think it was probably where operators have tried to price themselves out because they don’t want to deliver – they don’t have capacity or whatever – and it had set the pricing to £10,000/MWh and the National Grid algorithm had accepted it and it was relatively small volume.”

“It’s not the case of a huge scandal for billpayers but it led National Grid to pause that go back and tweak some of the coding.”

In an industry webinar in February, the ESO acknowledged the issue, saying in the “rare case” that the optimiser is fast ramping a unit either before or after a boundary between settlement periods this could result in high-cost bids or offers being accepted.

In response, the ESO suspended its use of the OBP for battery units until it was able to implement a fix for this defect on 8 January.

Since this hiccup, Mason says there has been “quite an improvement in volumes. I think there’s definitely positive movements in the right direction.”

According to figures from market intelligence firm Modo Energy, the ESO dispatched an average of 990MWh of bids and offers from batteries each day in December, with the figure rising to 1,110MWh in January. This compares to a daily average of just 200MWh in January last year.

As a result, Modo said average Balancing Mechanism revenues for batteries hit £540/MW during January this year, marking the first time these monthly revenues exceeded earnings from frequency response services, which were £210/MW lower following a drop off in prices.

Looking forward

Mason says: “There’s still quite a long way to go, but that’s going to come as the operators in the control room start to get more confidence in this”.

Like Guest, he believes the launch of the Balancing Reserve will help to bring this confidence: “I think one of the key issues that National Grid has at the moment around storage is not knowing how much storage is going to be available for several hours into the future.

“With traditional plant, they can be sure that the gas plant is going to have a capacity of say 500MW, no matter what, whereas with storage they know that can move around.

“If you’ve got a mechanic to actually contract and reserve volume at that capacity into the future, that should give a lot more confidence and therefore allow more volume to be to be allocated into the battery zone.”

Mason says the ESO has been trying to improve transparency, for example, by providing codes on each occasion to explain why a particular unit has been skipped. He says a lot of skips “went into a fairly opaque code so that wasn’t overly helpful,” but adds: “The less skips you have the less of an issue that that becomes so I think the fact that we’re getting more volume is inherently going to be helpful in that.”

“I’d rather focus on reducing the skips than improving the way they talk about it,” he summarises.

Bukhsh says the bulk dispatch optimiser will be just the first a “suite of tools” the platform will eventually feature: “There are further add-ons which are going to come in in future.”

Over 2024 and 2025, the ESO says plans to “incorporate a wider range of into the bulk dispatch process” and transfer reserve service over from the Ancillary Services Dispatch Platform. It says this and the Balancing Mechanism will be replaced by the OBP by 2027.

Bukhsk says optimising across all units will eventually present a computational challenge as their numbers increase: “I can’t say that 2000 units is the limit or 3000 units is the limit because it depends on some many other factors as well, but for these optimisation models, these mathematical models, there is a computational limit beyond which they’ll start to slow down significantly.”

“At the moment, it works just fine with two, three hundred units but if we were to scale it to the order of thousands, then that’s going to be a challenge, which we’ll see if the research can address in the coming years.”