IEA Wind TCP

IEA Task 47 Turbinia started in Spring 2021 after which the project got into full swing.

Within WP1  several countries started to define detailed aerodynamic measurement programs and more and more measurements are collected for a large variety of conditions (including storm conditions).

Also new measurement techniques are being developed, e.g. MEMS based pressure sensors , pressure belts, wake rakes, fibre optic sensors.

The main motivation of TURBINIA lies in the fact that detailed aerodynamic measurements are very specialized and the wind turbine society has very little experience with doing such experiments. In order to share the experiences on this specialized field of aerodynamic measurements a “recommendation document” is  made which describes several issues (some of them very practical) with respect to aerodynamic measurements. Much progress has been made on the chapters which describe the pressure measurements (including measurements with fibre optic sensors), non-dimensionalisation of loads and standstill measurements. This document is expected to help future experimentalists so that they donot need to invent the wheel again.

The document also addresses an important issue which hampers the full exploitation of these measurements and which is caused by the fact that the measurements (And the underlying machine data) can at the best be used for validation by the research party which carries out the measurements. They may however not share the data to third parties where previous IEA Tasks have shown how beneficial  such joined analysis of measurements is.

This is in particular true for measurements on large scale modern wind turbines. Fortunately data from smaller turbines can be shared. This includes recent measurements and a turbine description from OST on a 6 kW turbine.  https://mostwiedzy.pl/en/search/openResearchData?s=aerosense. Also aerodynamic field measurements from the previous IEA Task 14/18 projects have, together with their model descriptions, been uploaded on A database of detailed aerodynamic measurements on wind turbines, generated in TCP Wind IEA Task 14 and IEA Task 18 (zenodo.org)

Even though these are measurements on small turbines, they are still unique since they are (with the DanAero experiment) the only public aerodynamic measurements on wind turbines in atmospheric conditions.

Moreover much effort was spend on a calculational case around the DanAero experiment where the comparison (between calculations and measurements and between calculations mutually) is done on the basis of time series.  Such comparisons on the basis of times series are  challenging amongst others due to the fact that generating representative wind from the meteorological mast measurements is difficult in view of the relatively low resolution of meteorological sensors. Much potential is seen for time series comparisons with  the new experiments in TURBINIA which apply LIDAR. The much higher resolution of LIDAR wind speed measurements makes it possible to generate turbulent wind speed which will agree better with reality.

The comparison also showed differences between results from higher fidelity codes and results from lower fidelity BEM models which were attributed to difficulties in modelling non-uniformities with BEM. Thereto it should be known that the modelling of sheared conditions (and other flow non-uniformities over the rotor plane) inherently violate the assumptions in BEM.

For this reason a second round on the DAnAEro turbine was defined which consists of two sub cases at sheared conditions at different axial induction factors (a~0.38), and (a~0.25). Again calculations from lower fidelity BEM codes are compared with results from higher fidelity free vortex wake and CFD codes. In all cases the load amplitude from BEM is much higher than the amplitude from the corresponding higher fidelity model which implies a strong overprediction of fatigue loads from the industrial BEM models. Moreover a strong sensitivity is found on the way how the induced velocities are modelled in BEM.

The fact that BEM overpredict the load amplitudes is a result with very high impact in particular because these differences will increase with increased rotor size (The DanAero turbine has a diameter of only 80 meter). In order to understand the importance of this finding it should be realized that literature generally describes aerodynamic design models for industrial calculations as a  basic BEM model complemented with engineering add-ons (e.g. for modelling of 3D  or unsteady effects). Although it is generally acknowledged that there is wide variety of engineering add-ons which can have an impact on the results of a design code, the underlying BEM model is considered similar in every design code and it is not seen as a source for differences. This is now found to be a wrong assumption because loads as calculated from design codes are highly sensitive to the precise implementation of the BEM method. In other words THE BEM model does not exist and the resulting fatigue loads may differ considerably from code to code even when the same engineering add-ons are used.

In WP2 a first preparatory simulation case was defined for the IEA 15-MW Reference Wind Turbine which aims to check the correct aerodynamic and aero-elastic input so that trivial input errors are eliminated in the follow-up calculational rounds. Thereto, aerodynamic simulations are carried out under relatively simple conditions, with uniform steady inflow. These activities were coordinated with IEA Task 37.  The input for CFD codes was not available yet but it was generated in Task 47.

Eventually the agreement between results  was very reasonable. An important finding is the large torsion angle on the blade of such a 15 MW turbine, which is found to be in the order of 2 degrees at the tip even for a relatively low wind speed of 7.5 m/s. Such a large torsion angle obviously has a significant impact on the performance, the control and the loads which makes the accurate modelling of it very important and which implies that the relative large differences in calculational results needs to be understood.

At a meeting on May 22, 2023 the  project group assessed the calculational results from this preliminary case to be satisfactory after which turbulent calculations are performed. A CFD empty box was produced by DTU with a wind speed of (approximately) 7.5 m/s in the rotor plane in line with the wind speed from the preliminary case.

First scoping calculations were carried out by DTU and TNO with satisfactory results. Amongst others the turbulence intensity was low enough to avoid controller activity (which is important to prevent uncertainties from the controller model. In this way differences in results are solely caused by the aerodynamic-aero-elastic modelling).

The definition of the turbulent case has then been finalised after which it is distributed within the group so that they can start these calculations. This is complemented with a further assessment of the aero-elastic modelling similar to the analysis in Task 29, led by Polimi.

Several project meetings took place, mostly online but a personal meetings was held on May 30 2022 at TNO Delft, the day before the Science of Making Torque conference, in November 2022 at the University of Stuttgart and on May 22, 2023 the day before the Wind Energy Science Conference. Another personal meeting is planned in Firenze on May 28, the day before the Science of Making Torque confernece