Achieving alignment in energy production modelling

A recent Windpower Monthly webinar, produced in partnership with Vestas, discussed why the industry needs to converge to deliver greater accuracy in energy production analysis for wind projects

Accurate energy production is key to the industry moving forward
Accurate energy production is key to the industry moving forward

Overestimating and underestimating wind production can both be problematic and have a detrimental impact on project valuations, business models and project transactions further down the line – according to the speakers during a recent Windpower Monthly and Vestas webinar. 

How inaccurate energy predictions can impact projects

Overstating production can end up with different entities “taking haircuts” on project value, which becomes unstandardised, said Jamie Scurlock who is head of strategic procurement at RES. “When the corrections come through, it is effectively individual regions and individual investors who will take their own view and it becomes different to manage as an industry,” he says. 

“Underprediction may win on one or two projects because the investor gets a better deal on that project, but then the next project doesn’t get built because it doesn’t get assessed properly at what value it could bring,” he says. 

From the perspective of an investor, a project’s bankability can sometimes be addressed through price adjustment says Natalia Svejgaard, investment director at IFU. From experience, however, she says inaccurate energy predictions, often in combination with other factors, such as overstated wind scenarios, can have implications for mobilising senior debt, for debt sizing and for failing to meet completion deadlines set out in loan covenants. 

Accurate energy production is key to the industry moving forward, says Scurlock. “We need as accurate models as possible and harmony across the industry. The issue is due partly to the multiple sources of energy production estimates that feature as a project is developed.  

“If an ‘independent engineer’ is involved in an energy yield prediction and another is involved on the client side and another is involved on the investment side – if there is a disagreement there between those independents on how to deal with blockage, or wakes, or other issues, then the project value can change fairly rapidly over that space of time when really what the investor and the developer and all parties want is certainty during the period of transaction.”

Accounting for inaccuracy in a project’s business model 

Jean-Philippe Salomé, vice-president of industry at EDF Renewables, says that after 20 years of operation for EDF Renewables, the team there is aware of the importance of quality and accuracy of energy production assessments. Going forward this will be even more critical for wind as it competes against solar PV projects in technology agnostic auctions. 

In the case of a project or portfolio where the uncertainty is higher than the developer would normally expect, a bias, which can be as high as 5%, is applied to the adjustment factor/P50. “We can take a buffer on the production into account for when the uncertainty level is too high.”

“Sometimes it is the case where we have an uncertainty level that is lower our target. This can happen when we have flat terrain and high wind or an extremely good measurement campaign. In the case of when we perform a repowering we value the fact that we can rely on accurate long-term real production data from the site. This brings higher certainty and value to the project,” Salome says.

“Definitively, reliable energy estimates have an impact on the way we evaluate a project and finally on the business case,” he adds. 

Identifying factors that lead to energy prediction bias 

Factors that may be contributing to large variance within the industry in relation to energy prediction bias and loss factor calculations can be attributed partly to the trend towards taller turbines with larger swept areas, designed to capture more wind resource. Scurlock likens the challenge to using an instrument the size of a tennis ball to measure and predict a swept area the size of three football fields, to determine how the turbine will perform, which can be further influenced by variations in wind speeds and extreme weather conditions. 

The methods for monitoring and measuring that were appropriate five to 10 years ago, may not be today says Scurlock who notes increased use of Lidar/remote sensing as an important addition to meteorological masts for campaigns, for establishing more accurate measurements. But newer developments can also lead to the challenge of harmonising methods and approaches across the industry. 

The variation the industry is witnessing when it comes to interpreting the data gathered from measurement campaigns to inform energy predictions, can result in a higher cost of capital, according to Svejgaard.

Requirement for higher levels of energy predictions accuracy

Improved modelling and reporting of uncertainty is a priority, says Svejgaard. In the case of a solar PV project, if the expected performance ratio cannot be achieved by the time the project reaches commercial operation date (CoD) more modules can be added. “But in a wind project it is not just a matter of adding another turbine or two – it’s not really a likely scenario.” 

She says improved modelling and reporting of uncertainty is a priority. “We’d like to see accurate evidence-based uncertainty models potentially based on time-based energy prediction approaches using risk assessments, to identify, report and potentially mitigate technical risks… as well as improved site measurement tools and more accurate wind flow and wake models.”

Progress towards harmonising energy predictions 

Good examples of efforts to harmonise and standardise in relation to energy prediction, include the Power Curve Working Group and also the work of the Consortium for Advancement of Remote Sensing (CFARS), for helping to guide the industry on the advantages and also the shortcomings when deploying the technology.

“CFARS is a great example of collaboration, which has shown some great results so far and will continue to develop and offer the industry better insights and better ways of working. It is a good example of an industry group coming out of the people who have their interests aligned,” says Dr Chris Ziesler, director advisory services, North America UL Renewables, 

He says in terms of standardisation the industry has made progress, according to Ziesler. “As technical advisers our role is to try to be consistent and give everybody a good read.”  

He refers to the IEC-15 Committee, which is bringing “more refinement and precision” and is a good example of standardisation and collaboration. “It gives us a helpful and common framework to talk about loss factors, in particular. The majority of people now understand what you have to do to get good data in terms of met campaigns but there is still a lot of discussion around loss factors and a lot of arguments around these, so having a common framework is important.”      

Paul Leask, service line leader for project development and analytics, Energy Systems DNV wants to see better or increased understanding of remote sensing measurement uncertainty. “These devices are being used more by developers and as an industry we need to support and facilitate that if that’s the way developers want to go.”

DNV advocates is also in favour of a common definition of a validation or how to conduct one so that it provides a like for like comparison for different methodologies. “This would remove the noise around the competitive environment,” he says. 

Bridging power curve measurement and real-world turbine performance 

Another challenge is reducing the gap between power curve measurement standards and turbine performance in real world conditions. To help address this, Vestas has released work on achieving a track record on power curve predictability, through being more transparent about its power curve prediction modelling and methods.

Vestas senior specialist in product performance and warranties Michael Pram Nielsen says: “Practically speaking this has led to a reduction in energy prediction loss factors.” This was achieved through the OEM sharing more than 300 IEC-compliant performance test results for auditing. “We have also seen the positive impact on several wind farm construction projects,” he says.  

Pram Nielsen says  that alignment on energy prediction risk factors that can lead to various revenue losses is important going forward. “Maybe year-to-year variation is less important but we need to improve our understanding of what is causing this variation. We need more industry standardisation around uncertainty and loss variation methods.” 

By building trust across different parts of the renewable industry energy advisers can establish trust in models that OEMs are working with, he says. 

Click here to watch the webinar in full.

Have you registered with us yet?

Register now to enjoy more articles
and free email bulletins.

Sign up now
Already registered?
Sign in

Partner content