'The wind power industry needs to leave many of the old metrics behind – AEP is passé'

Daniel Luecht, CEO of new artificial intelligence (AI) software start-up Nash Renewables, is pulling no punches. “There is no real value in building another wind farm in northern Germany – or anywhere else – that functions the same way or is designed the same as the others before it, from a financial or indeed decarbonisation perspective,” he says.

“Countries are clearly on a trajectory to reduce or remove [subsidy] support [for clean-energy projects] and this is moving the energy transition into a different phase in a KPI [key performance indicators] sense, if you like. Leaving AEP behind is going to be quite tough – we are all so accustomed to it,” says Luecht.

“And we see with the behaviour of the big utilities, in the connections between the people developing new renewable assets and the ones trading them, there is still a missing link and that is what we are trying to help to solve.”

Lauenstein adds: “The wind industry is in the process of transitioning into the post-subsidy market. The era of government-sponsored fix-price remuneration schemes is increasingly over. Even the remaining auction schemes often entail significant exposure to market volatility.”

The Nash view is that by “bringing the technology ecosystem and the electricity market ecosystem together” the wind industry can – and indeed must, if Paris climate targets are to be reached – make the leap beyond AEP and its “passé” companion metric, levellised cost of energy (LCOE), as wind takes up an increasingly dominant place in global energy production capacity.

“Today these ecosystems function in isolation, they don’t ‘talk’ to each other, and in fairness they are both hugely complicated systems,” says Lauenstein.

“We believe that in many cases, this is resulting in the industry’s focus on AEP and LCOE and not what actually matters: ‘value’ of energy. And in doing so, deploying the wrong or, at least, ‘un-optimised’ assets, both for the investors and for decarbonising the grid.”

Based on a set of AI-modelled wind farm case studies that factored in a range of “predictive” datastreams, Nash has put to the test its theory that “a far better way” of planning where to build wind farms, how to design them, and “when exactly” to run the turbines would be one aligned to minute-by-minute data points “deeply integrating asset technology configuration and electricity in the market optimisation and in your decision-making process”.

The results from running the Nash AI software – which mashes everything from data-feeds on weather and energy markets, right down to individual turbine fatigue loads – on three operating wind farms point to the potential for “positive disruption” at a crucial stage of the wind industry’s evolution, where the “unwavering concentration on cost, cost, cost is killing the supply chain”, says Lauenstein.

On one of the Nash-modelled sites, in northern Germany, though annual wind farm uptime was 25% lower that in the previous year – 76% rather than 98% – and AEP slipped 3%, by “correlating wind” – tailoring production levels onto the grid based on highest possible capture price – revenue from the wind farm jumped 15% and carbon displacement 2%.

A site in western Denmark saw 8% higher revenue from 9% lower AEP and one in Scotland an 7% uptick in revenue versus a 2%-lower AEP, once switched over to “smart delivery” of power to the grid based on real-time and forecast price.

Wind power’s ‘incremental value’

“Many mature wind countries have nearly ‘enough’ installed wind capacity to power their grid when the wind is blowing strongly,” says Luecht. “What we need to do now is to deploy new wind farms with a focus on the incremental value they provide, not how much energy they can produce in a year, to have our wind farms together orchestrated in a such a way that they create stability in the grid.

“How mad is it to try to capture all the wind you can on a wind farm and send it onto the grid even when the market has power at a close-to-zero or even a negative price point? Set an annual target for ROI [return on investment] and then run the wind farm to achieve or overachieve that day after day using an AI system.”

Lauenstein adds: “This is automatically what you come to when you start to look at deeply integrating asset technology configuration and electricity in the market optimisation and into your decision-making process.

“The volatility of the market is not going to go away because you have energy storage – though this is what OEMs generally believe. Energy market prices will determine the profitability of wind assets. With that, these assets become subject to the market’s price volatility and cannibalisation.

“And market volatility is not going to go away because you have energy storage or P2X [power-to-X] – nearly every scientific study or commercial forecast confirms that. ”

Top-line, according to Nash: the wind industry could be getting up-to-15% higher risk-adjusted revenue from post-subsidy projects if the “time resolution” on a project was tuned to the minute- or hour-level – that, for a medium-sized onshore farm would translate into €10m ($11.3m) in added investment value.

“Today wind asset business cases are mostly built on an annual time resolution: AEP multiplied by average market capture price results in projected revenue. Bring this down to the minute- or hour-level [for a wind farm of the future] and you’ll get a very different perspective on what asset choices unlock the most value,” says Lauenstein.

“If you break down a 25-year design lifetime into 220,000 hours and start looking at the unique relationship between revenue capture and asset degradation in each hour, you start noticing that unconventional asset choices may yield much better investment and decarbonisation value.”

Daniel Luecht adds: “If you have blinders on, if you are saying ‘I need to run this wind plant 24/7 and that is that’, then you are not even looking at the choices that could be available to you. But remove the constraint that AEP may represent and suddenly that are myriad choices open to you: taller towers, bigger rotors, different wind farm layouts – all of which can make an asset much more profitable.

Annual energy production ‘a constraint’

“This [asset optimisation] is hardware-driven in a sense, but in a much bigger sense it is software-driven, it is adaptive.”

The response from energy traders at “one of Europe’s largest utilities” to the Nash model, says Lauenstein, was “very positive”.

“As they told us, looking at an asset this way, it means a wind farm for them becomes [in one regard] like a coal- or gas-fired power plant because you can turn it on or off based on the market price.”

There is also the benefit of thinking about wind power production in “life of plant” terms, he adds: “Think of 2020, a ‘lost year’ for European wind because the prices were so low, everyone missed their business case.

“But an asset could be ‘designed’ to run at lower capacity in these sorts of periods and then – in 2021, for instance, run it at full capacity when prices are super-high. Uptime is averaged over life of plant, so why not decide to ramp your production up or down when it is most profitable.

Winning in a ‘lost year’ like 2020

“In five years, we have no doubt this value-focused approach will be the standard. It is one of those rare occasions where paradigm shift will be beneficial for almost all parts of the renewable energy ecosystem,” says Lauenstein.

“But right now, of course, we see the potential for resistance. We are a start-up and this is a disruptive model.”

The higher calling for Nash, underlines Luecht, is climate action. Decarbonisation has been a “side effect” of the energy transition to this point, he says, but will become “hugely important to the true ‘value’ of wind power to the grid” as the world weans itself off fossil fuels in the coming decades.

“As energy production moves to hourly resolution from monthly and yearly, it will increasingly matter when you produce with your wind asset. You’ll get many more carbon credits if you produce energy in times of high carbon intensity in the grid and there is an abundance of renewables on the grid pushing out the fossil fuels.

“Hourly or minute-to-minute carbon accounting will be the new financial accounting.”

Luecht sees the rise and rise of corporate renewable power production agreements (PPAs) – which according to Bloomberg NEF figures last year shot up 24% to over 30GW – as “playing in favour” of the Nash approach. “There will be PPA originators that can buy into our technology and use the algorithm for the agreement. Could the adoption of corporate PPAs speed the global build-out of renewables? 100%.”

This way, he notes, “producing assets would all be playing into the grid on a demand basis – as primary energy supply, as a grid service – rather than just whenever the wind blows”.

Lauenstein stresses: “Decarbonising the grid at the speed we need to requires a renewable energy supply chain that functions based on ‘value’ not volume and cost. This is going to be a massive transformation.

“And,” adds Luecht, “it is unavoidable. It is only a matter of how fast we can – and need – to move. Because renewables need to take over from all other forms of energy production currently in operation. Nothing less.”