Influence of wind conditions on wind turbine loads and measurement of turbulence using lidars

Abstract

Variations in wind conditions influence the loads on wind turbines significantly. In order to determine these loads it is important that the external conditions are well understood. Wind lidars are well developed nowadays to measure wind profiles upwards from the surface. But how turbulence can be measured using lidars has not yet been investigated. This PhD thesis deals with the influence of variations in wind conditions on the wind turbine loads as well as with the determination of wind conditions using wind lidars. Part I of the thesis focuses on analysis of diabatic wind profiles, turbulence, and their influence on wind turbine loads. The diabatic wind profiles are analyzed using the measurements from two offshore sites, one in the Dutch North Sea, and the other in the Danish North Sea. Two wind profile models are compared, one that is strictly valid in the atmospheric surface layer, and the other that is valid for the entire boundary layer. The second model is much more complicated in comparison to the first. It is demonstrated that at heights more than 50 m above the surface, where modern wind turbines usually operate, it is advisable to use a wind profile model that is valid in the entire boundary layer. The influence of diabatic wind profiles under steady winds on the fatigue damage at the blade root is also demonstrated using the aero-elastic simulation tool Bladed. Furthermore, detailed analysis of the combined influence of diabatic wind profile and turbulence on the blade root flap-wise and edge-wise moments, tower base fore-aft moment, and the rotor bending moments at the hub is carried out using the aero-elastic simulation tool HAWC2. It is found that the tower base fore-aft moment is influenced by diabatic turbulence and a rotor bending moment at the hub is influenced by diabatic wind profiles. The blade root loads are influenced by diabatic wind profiles and turbulence, which results in averaging of the loads, i.e. the calculated blade loads using diabatic wind conditions and those calculated using neutral wind conditions are approximately the same. The importance of obtaining a site-specific wind speed and stability distribution is also emphasized since it has a direct influence on wind turbine loads. In comparison with the IEC standards, which generalize the wind conditions according to certain classes of wind speeds, the site-specific wind conditions are demonstrated to give significantly lower fatigue loads. There is thus a potential in reducing wind turbine costs if site-specific wind conditions are obtained. In this regard we then are faced with measurement challenges. The current industry standard for the measurement of wind speed is either the cup or the sonic anemometer. Both instruments require a meteorological mast to be mounted at the measurement site. For measuring the wind profile the instruments need to be mounted at several heights on the mast. To install a mast and set up these instruments is quite expensive, especially at offshore sites, where the cost of foundation increases significantly. Besides, there are problems with the flow distortion that have to be taken care of. In order to overcome these problems it would be ideal to have a remote sensing instrument that measures wind speed. Wind lidars are capable of doing that albeit with a price. Part II of the thesis deals with detailed investigations of the ability of wind lidars to perform turulence measurements. Modelling of the systematic errors in turbulence measurements is carried out using basic principles. Two mechanisms are identified that cause these systematic errors. One is the averaging effect due to the large sample volume in which lidars measure wind speeds, and the other is the contribution of all components of the Reynolds stress tensor. Modelling of turbulence spectra as measured by a scanning pulsed wind lidar is also carried out. We now understand in detail the distribution of turbulent energy at various wavenumbers, when a pulsed wind lidar measures turbulence. The lidar turbulence models have been verified with the measurements at different heights and under different atmospheric stabilities. Finally, a new method is investigated that in principle makes turbulence measurements by lidars possible. The so-called six beam method uses six lidar beams to avoid the contamination by all components of the Reynolds stress tensor. The theoretical calculations carried out demonstrates the potential of this method. In order to avoid averaging due to volume sampling, a different analysis method is required, which has not been investigated in this thesis. To summarize the entire thesis, it can be said that more work is required to ascertain the influence of atmospheric stability on wind turbine loads. In particular, comparing with the load measurements will go a long way in consolidating the understanding gained from the analysis in this thesis. If lidars are able to measure turbulence, there is a tremendous potential for performing site-specific wind turbine design and making the class based design of the IEC standards obsolete.