The Effect of Wake Models, and Environmental Conditions on Wind Farm Layout Optimization

The Effect of Wake Models, and Environmental Conditions on Wind Farm Layout Optimization

Daneshbodi, Amir (TU Delft Electrical Engineering, Mathematics and Computer Science)

Zaayer, M B (mentor)

Delft University of Technology

Aerospace Engineering | Aerodynamics and Wind Energy

2018-02-19

The purpose of this research is to broaden current knowledge of Wind Farm Layout Optimization (WFLO). Until now, there are many research conducted on different optimization methods, annual energy yield calculation processes, and wake modeling methods. But, few papers have addressed the effect of different modeling parameters on WFLO in a form of sensitivity analysis. The results would be helpful in order to have a better overview on the effect and importance of each modeling parameters on the objective function calculation which would finally affect the layout found by the optimizer. The first phase of the project defines and categorize parameters affecting the WFLO. The second phase involves setting up the methods to find the effect of the categorized parameters such as turbulence intensity, wind direction sampling step, and wake model selection on WFLO.
The research is done on the analyzer stage and ignored the optimizer. The analyzer has function of calculating the annual energy yield and providing the result to the optimizer module for the WFLO process. Moreover, the study starts from a single wake calculation, to wind farm power production for an individual wind direction, and finally to annual energy yield production. Following these steps is helpful because the same steps happen in the analyzer, and it enables us to determine effects of the parameters within the flow of calculation in the analyzer. These steps are vital to sheds light on the origin of each effect that happens on the annual energy production level and helps to understand if they are magnifying or disappearing during these steps. This study shows that Jensen, Larsen, and Ainslie wake models are similar in ranking the annual energy yield of a wind farm with random configuration, although they have different behaviours in increasing the annual energy production by increasing the spacing in a wind farm (which is more sensible in relatively lower spacing). For instance, in a wind farm with turbines relatively close to each other, the Larsen model has the most increase due to the increase in spacing, which leads the analyzer to find a larger wind farm as the solution (compared to Ainslie and Jensen) when costs (like cabling costs) are also taken into account combined with the annual energy production.

Wind Energy
Wind farm layout optimization
Wake model
Annual energy production

http://resolver.tudelft.nl/uuid:25af606d-bd36-43ae-9f0f-b1090e272137

Student theses

master thesis

© 2018 Amir Daneshbodi