Specific aerodynamic measurements are needed to understand the very large challenges of the aerodynamics of large wind turbines in atmospheric flow. These measurements only can capture the details of the flow physics.

Detailed aerodynamic measurements are even an absolute necessity when validating wind turbine design codes since a comparison with conventional measurements of global (rotor) blade loads is far too crude and it does not provide the accuracy of such codes. This is due to the fact that a good agreement between calculated and measured rotor (blade) loads is often caused by compensating errors (e.g. an overprediction at the root can be compensated by an  underprediction at the tip or an overprediction in lift can be compensated by an overprediction in drag). In such cases a comparison with global load measurement yields a good agreement and the very misleading conclusion that the code is accurate.

An example is given in figure 7 from IEA Task 18. It shows the ratios between the rotor shaft torque calculated by TNO (ECN at that time) in comparison to measurements from the NREL Phase IV experiment as function of mean wind speed. A ratio of 1 implies a perfect agreement (i.e. the target value). A ratio > 1 implies an overprediction and a ratio < 1 implies an underprediction. The comparison is generally very good with the largest discrepancy being 10% at  roughly 13 m/s and discrepancies at other wind speed which are often 5% or even less. Hence in the conventional case where rotor shaft torque (or power) measurements would be available only, the conclusion is that the code is performing well.

However the measurement of pressure distribution at 5 radial positions as available in the NREL experiment offers a much more detailed validation see figure 8. This figure shows the comparison of sectional forces in normal and tangential direction. It is then very striking to see very large discrepancies (tens of percents) between calculations and measurements in particular at the higher velocities. This shows that the good agreement between calculated and measured rotorshaft torques is a result of compensating errors (which are explained in [7]) and the quality of the calculational code was not good at all! This clearly proves that validation and improvement of aerodynamic models needs detailed aerodynamic measurements and cannot be based on global loads.

Mexnext stands for the follow-up (NEXT) of the MEXico project. It is the acronym used in the first three phases of Task 29, in which the results from the Mexico project were analysed. In the Mexico project detailed aerodynamic measurements were carried out on a 4.5 meter diameter wind turbine placed in the large German Dutch Wind Tunnel DNW. The Mexico project stopped after the measurements were collected. The analysis of data took place in Task 29.

A first benefit which should be mentioned is the fact that an IEA Task offers a platform to share experiences and knowledge on the very specialised field of aerodynamic modelling and measurements. All participating bodies in Task 29 acknowledge that this leads to huge mutual benefits with enhanced knowledge levels at all parties.
These shared insights and analyses from the Task partners improved the quality of the various experimental databases after which these databases could be provided as high level validation material for design codes to a countless number of partners in the IEA Tasks but also to third parties all over the world. This holds for the field measurements taken in IEA Tasks 14/18, the NASA-Ames wind tunnel measurements from Task 20, the Mexico wind tunnel measurements from Task 29 and the DanAero field measurements from Task 29. The resulting knowledge of uncertainties in design calculations from these validations help designers to assess the validity of their design calculations, and allows standardisation bodies to recalibrate safety factors in design standards. The measurements also led to fundamental model insights and aerodynamic understanding which clearly supported the improvement  of aerodynamic wind turbine models as implemented in design codes. This is illustrated by the proposition  from [7]  (i.e. a PhD thesis based on the results of IEA Task 14/18/20 and the first phase of Task 29)
“nowadays there isn’t a designer to find who would dare to design a wind turbine with the very basic aerodynamic modelling as applied in 1980’ [15)]”

More specific figure 9 illustrates how knowledge on aerodynamics, together with knowledge on other disciplines, play a role in assessing the cost of energy.

The Levelised Cost of Energy (LCOE) of wind energy is largely determined by the performance and capital costs, where capital costs are directly related to turbine loading. Both performance and loading are predicted with higher accuracy from the improved aerodynamic models. The cost model of FP7 INNWIND.EU shows that a combination of 5% increase in energy yield with a 12% decrease of the design loads of the turbine critical components and offshore substructure lead to a 10% LCoE reduction (conservative estimate). It is also important to realise that recommendations on optimal blade design and blade innovations, e.g. the optimal choice of airfoils or application of blade add-ons are important means for reducing the LCoE. These optimisations and innovations can only be succesful if the blade aerodynamics is understood.