Wind energy: New operating strategy takes into account
variable production costs and electricity prices for higher

It is possible to achieve maximum returns in the direct marketing of wind power if plant control systems consider not only fluctuating electricity prices but also ever-changing production costs (i.e., wear and tear of plant technology). As part of the “KORVA” research project, the Fraunhofer IEE and its partners have now developed the very first model for helping system operators to incorporate such considerations into their control strategies. This software tool enables them to control the plants so that they strike the optimum balance between revenue and costs. Machine learning (ML) methods are also used for this purpose.

“Wind turbines provide investors with the most profit when they are run at optimum efficiency, maintaining the right balance between revenue and costs. This is where KORVA comes into play: By significantly extending the service life and achieving higher energy yields, it makes the entire wind farm project considerably more profitable,” explains Dr. Boris Fischer, KORVA project manager at Fraunhofer IEE. “However, this is an extremely complex task. This is because system operators need certain information that is usually unavailable to them in order to achieve this level of efficiency. Our research project has bridged the gap. We are therefore breaking new ground for the industry and, in doing so, helping to make wind power an even more attractive investment opportunity.”

Fraunhofer IEE was responsible for managing and coordinating KORVA. The following partners were also involved in the research project: Nordex Group (manufacturer), ABO Wind (operator), Steag and Statkraft (direct sellers), 8.2 (independent expert) and TÜV Süd (certification body).  The three-year project, funded by the German Federal Ministry for Economic Affairs and Climate Action, was successfully completed at the end of April 2022. The end result was optimized operating strategies that are more environmentally and economically viable compared to conventional ones. In specific terms, their application resulted in the service life being extended from 20.7 to 30.4 years in one case study. The wind farm produced 122% of the energy that it would have been able to generate without load-dependent optimization until the end of its working life. The additional yield in relation to the pure rate of return increased by 312%.

Knowledge of revenue and costs

The production costs of wind turbines largely depend on how much wear and tear their components go through during operation. This is not constant over their service life but varies, primarily depending on wind conditions. In addition to wind speed and direction, as well as shear, the extent to which the individual turbines are exposed to turbulence is a key factor in this respect. As a result, a higher revenue on the power exchange does not necessarily mean a higher profit. If, for example, the production costs are equally high during this period, the system operation might be less economical than at times when the revenue is lower but the production costs are also very low. In such cases, it might even make economic sense to temporarily take the turbines offline to prolong their service life — KORVA has shown that the additional, higher revenue gained during the extended service life more than offsets any lost revenue.

Such decisions cannot be made without knowing both the current and expected revenue on the power exchange as well as the production costs at the relevant times. Operators can use this knowledge to design and adapt their control strategies so that profit is maximized over the service life of the turbines. Turbines can also supply more energy with regard to their total service life, as they are run less frequently at times when the likelihood of wear and tear is very high. The project demonstrated the revenue increases that can be expected in practice under given boundary conditions.

Optimized operational management for maximum returns

In the KORVA project, Fraunhofer IEE worked with its partners to develop the first optimization tool for operational management that meets these requirements. The new software generates generalized schedules for wind farms, which are derived from the expected revenue and costs. The calculations are based on factors such as data on wind conditions, electricity market prices and operating costs as well as on wear-and-tear models. In a final step, these generalized schedules are integrated into operational management via newly developed data interfaces.

During the tool’s development, the Fraunhofer researchers used meteorological data from four wind farm regions within Germany to map different location characteristics. The selection criteria were high- and low-wind sites and sensitivity to the wind power market value. Given the volatile nature of the electricity market and the changeability of wind conditions, the estimates for the expected loads and revenue of individual turbines must be made available very quickly. Fraunhofer IEE therefore incorporated sophisticated algorithms from the field of machine learning into the new optimization tool to help perform these tasks — a computationally efficient alternative to mechatronic models or commercial software. More specifically, the scientists deployed artificial neural networks. These provide quick yet sufficiently accurate estimates of the expected loads and yield for individual wind turbines.

“Our optimization tool enables system operators to adjust operational management on an ongoing basis so that the right balance between energy output and component wear and tear can be maintained to achieve maximum total returns. This ensures that the investment in the wind park is more profitable over its entire service life. Looking ahead to the future, other systems with load-dependent aging effects such as batteries or electrolyzers should also be able to benefit from the methods developed in KORVA,” explains Fischer, Fraunhofer IEE.


A presentation of the project results will be held online on December 16, 2022 from 1-3pm. Registration via the link below or here. The event will be held in German.

More Information

Last modified: