The operators guard the capacity utilization of their wind farms like a state secret. The NZZ has now calculated it itself. The results are sobering.
When is a plant economical?
A high level of utilization is a basic requirement for plants to be able to be operated economically – especially in countries with a strong market economy, such as the USA. In Germany, there are other factors that affect economic viability, such as the type and amount of government funding and how it interacts with prices on the electricity market. Because Germany is helping the so-called reference yield model expansion even in poor locations. According to a calculation by the German Wind Energy Institute from 2003, the profitability of large systems in Germany starts at a capacity utilization of 23%. At that time, however, the operators received a very high fixed feed-in tariff from the state. With today’s market premium model, the head of the Stuttgart Chair for Wind Energy, Po Wen Cheng, considers a guideline value of around 30% to be realistic. Project developers of wind farms did not want to comment on the NZZ request.
How was the utilization of the systems calculated?
First, the NZZ calculated the amount of electricity produced hourly by 18,000 onshore turbines over a period of ten years using the programming interface of the Online Tools Renewables.ninja simulated. The capacity factor was then calculated based on the rated output of the turbines. Renewables.ninja is based on a model by Stefan Pfenninger from TU Delft and Iain Staffell from Imperial College London. The simulation of the average amount of electricity produced is based on weather data from NASA from the years 2011 up to and including 2020.
How precise is the occupancy rate calculated by the NZZ?
The mean deviation (RMSE) of the modeled utilization from the actual utilization is three percentage points on average over a ten-year period. As a rule, the model slightly overestimates the utilization, i.e. in reality the utilization of most wind turbines is even worse. This is plausible insofar as in the simulation shutdowns and throttling due to icing or noise and environmental regulations are not taken into account; these have a negative effect on the actual capacity utilization. For the comparison, the NZZ inquired about the annual amounts of electricity produced by a total of ten wind farms of different sizes and distributed throughout Germany, for which the municipal operator was obliged to provide information. The model underestimated locations where individual wind turbines are on a hill in an otherwise flat landscape, such as in front of Stuttgart or in Munich. These wind turbines were removed from the data set.
Why aren’t all the wind turbines shown on the map?
The NZZ calculation is based on a data record from the Federal Network Agency. According to this, there are currently around 28,000 wind turbines on land in Germany with a nominal output of 100 kW or more. However, the dataset was often incorrectly filled: for example, with incorrect geographic coordinates, imprecise information on the turbine type, or with rotor diameters that do not match the specified turbine type. In addition, the model used by the NZZ lacks technical data for some turbine types. All of this meant that after the data cleansing, 18,000 of the total of 28,000 systems remained. According to this, two thirds of the larger onshore systems are included in the data set.
Why does the text refer to individual turbines and not to wind farms?
The NZZ algorithm does not take wind farms into account, but known turbine types: with the same tower height and the same rotor diameter at a similar location – regardless of the wind farm operator. As a rule, therefore, no statement can be made about wind farms, as these often consist of different turbine types whose technical data may not be stored in the model.