Due to inhomogeneous and unsteady incoming wind conditions, it is expected that wind turbine airfoils under real operating conditions may encounter angle of attack variations of around ± 5° inducing load fluctuations to mitigate. It is then necessary to optimize the wind energy efficiency and the rotor lifetime by developing innovative control concepts with the intention of reducing load fluctuations on blades.

Strategies of circulation control acting at the blade airfoil trailing edge are usually investigated to allow lift increase and decrease. For aerodynamic bodies they are traditionally implemented by means of concepts including shape change, flaps, blowing, suction, etc., and often make use of the Coanda effect that keeps a tangential jet attached over a curved surface. This work aims at investigating the potentialities of surface plasma actuators, system identification and control theory to solve the circulation control problem and reduce load fluctuations. Such actuators are indeed characterized by quite low momentum coefficients and one of their main advantageous features is their short time response. As it can be assumed that the flow induced by the Dielectric Barrier Discharge (DBD) actuator behaves like a two-dimensional wall jet, it allows to perform a bi-dimensional action on the flow. The baseline aerodynamic configuration on the airfoil is a fully attached flow.

 Firstly, the objective of this work was to assess the potentialities of such a circulation control technique by conducting experimental testings. Surface pressure, load and velocity field measurements in the large subsonic wind tunnel of PRISME laboratory were performed to evaluate the actuator effectiveness and to highlight the flow mechanisms induced by the actuation for lift increase and decrease configurations. The maximum lift coefficient variation obtained for a chord Reynolds number of 2. 105 was ±0.1 depending on electrical operating parameters chosen for powering the plasma actuator. Secondly, system identification theory was used to build a model of the systems. The chosen input variable was the high voltage amplitude and the chosen output variables were the surface pressure levels obtained with 20 pressure taps distributed along the chord that permit to characterize the lift force variation. Different test cases were conducted by varying high voltage amplitude at a given angle of attack or by varying angle of attacks by impulsive or sinusoidal movement schemes for example. Results permitted to extract some involved time scales and to identify the system around some operating point. Interestingly, the system was found to behave like a piecewise linear systems when using wisely chosen pressure taps. This model allows then to correctly predict the lift force in controlled cases. This paper focuses on the analysis of such results.