Search

A wind farm optimization framework is presented in detail and demonstrated on two test cases: 1) Middelgrunden and 2) Stags Holt/Coldham. A detailed flow model describing the instationary flow within a wind farm is used together with an aeroelastic model to determine production and fatigue loading of wind farm wind turbines. Based on generic load cases, the wind farm production and fatigue evaluations are subsequently condensed in a large pre-calculated database for rapid calculation of lifetime equivalent loads and energy production in the optimization loop.. The objective function defining the optimization problem includes elements as energy production, turbine degradation, operation and maintenance costs, electrical grid costs and foundation costs. The objective function is optimized using a dedicated multi fidelity approach with the locations of individual turbines in the wind farm spanning the design space. . The results are over all satisfying and are giving some interesting insights on the pros and cons of the design choices. They show in particular that the inclusion of the fatigue loads costs give rise to some additional details in comparison with pure power based optimization. The Middelgrunden test case resulted in an improvement of the financial balance of 2.1 M€ originating from a very large increase in the energy production value of 9.3 M€ mainly counterbalanced by increased electrical grid costs. The Stags Holt/Coldham test case resulted in an improvement of the financial balance of 3.1 M€.&nbsp

Turbulence time series are needed for wind turbine load simulation. The multivariate Fourier simulation method often used for this purpose is extended for inhomogeneous and non-stationary processes of general probability distribution. This includes optional conditional simulation matching simulated series to field measurements at selected points. A probability model for the application of turbine wind loads is discussed, and finally the technique for non-stationary processes is illustrated by turbulence simulation during a front passage.

A coherent wind speed and wind direction change (ECD) load case is defined in the wind turbine standard. This load case is an essential extreme load case that e.g. may be design driving for flap defection of active stall controlled wind turbines. The present analysis identifies statistically the magnitudes of a joint gust event defined by a simultaneously wind speed- and direction change in order to obtain an indication of the validity of the magnitudes specified in the IEC code. The analysis relates to pre-specified recurrence periods and is based on full-scale wind field measurements. The wind speed gust amplitude, occurring simultaneously with a wind direction change, corresponds well to the recommended ECD value of 15 m/s, except for the complex terrain case, where estimated extreme wind speed gust amplitudes are seen to exceed the IEC value with approximately 50%. The estimated extreme wind direction gust amplitudes associated with the investigated European sites are low compared to the recommended IEC- values. However, these values, as function of the mean wind speed, are difficult to validate thoroughly due to the limited number of fully correlated measurements.

The vast majority of wind turbines are today erected in wind farms. As a consequence, wake generated loads are becoming more and more important. We present a new experimental technique to measure the instantaneous wake deficit directly, thus allowing us to quantify the wake meandering as well as the instantaneous wake expansion expressed in a meandering frame of reference. The experimental results are subsequently used in a preliminary verification of the basic conjecture of a wake meandering model that essentially considers the wake as a passive tracer.

This study reports structural dynamic characteristics obtained experimentally from an extensive testing campaign on a 34m long wind turbine blade mounted on a test-rig under laboratory conditions. Further, these experimental results have been compared with analog numerical results obtained from a very detailed FE model of the same blade using 3D solid elements. Both an undamaged and a damaged blade are investigated, and it is observed that the natural frequencies of the first few modes of the blade change very little due to a significant artificial damage imposed in trailing edge, whereas the mode shapes - especially if decomposed into the flapwise, edgewise and torsional components - contain information which might be helpful for detecting and localizing wind turbine blade damages.

Current state of the art wind power plant (WPP) controllers operate wind turbines (WTs) independently as individual machines, thus dispatching the WTs’ set points in an equal manner to all of them. To achieve optimal WPP control, three aspects need to be addressed: • Maximizing the yield (power production) balanced against turbine mechanical loading and electricity price • Enhancing WPP capability to provide ancillary services (primary, secondary, and tertiary reserves), and • Reducing operating costs (i.e. reduced fatigue load degradation of WTs and O&M requirement) over the lifetime of the WPP. The goal of TotalControl is to move the WPP controller design philosophy from individual optimization of WT operation to a coordinated optimization of the overall WPP performance. The TotalControl project aims to achieve this by developing and validating advanced integrated WPP/WT control schemes conditioned on grid demands and wind turbine fatigue damage limits. For developing and testing of the different WPP controllers, a range of high-fidelity and medium-fidelity simulation models are used. These models are already available in the consortium, but will be thoroughly validated against full-scale measurements in the Lillgrund WPP. Due to the complexity and multi-scale nature of WPP flow dynamics, the high-fidelity CFD-based models are very expensive in simulation time, e.g. requiring supercomputing, and therefore not well suited as control design models.

Quantifying accurately wind turbine wakes is a key aspect of wind farm economics in large wind farms. This research compares three engineering wake models with power production data from the Horns Rev and Lillgrund offshore wind farms. Single and multiple wake cases are investigated to verify the performance of the models in different conditions. The simulations reveal that the three wake models have similar behaviours for both wind farms although the turbine spacing and the turbulence intensity are different. The results prove the robustness of the models to provide accurate power predictions when the simulations are averaged over wind direction sectors of 30. However, all models significantly underpredict the power production of a single row of wind turbines using narrow sectors of 3 or 5. This discrepancy is discussed and justified by the wind direction uncertainty included in the datasets.