Modelling and recursive power control of horizontal variable speed wind turbines H.P Wang*, A. Pintea**, N. Christov***, P. Borne****, D. Popescu** * Nanjing University of Science and Technology, Automation School, China (Corresponding author: hp.wang@njust.edu.cn) ** University “Politehnica” of Bucharest, Splaiul Independentei, nr. 313, Romania (andreea.pintea@gmail.com, popescu_upb@yahoo.com) *** Lille University of Science and Technology, 59651, Villeneuve d’Ascq, France (nicolai.christov@univ-lille1.fr) **** Ecole Centrale de Lille, 59651, Villeneuve d’Ascq, France (pierre.borne@ec-lille.fr) Abstract: The paper deals with the modeling and power control of horizontal variable speed wind turbines using a new recursive controller. The problem considered is to realize a multi-objective control scheme covering the entire wind turbine working domain. The proposed recursive model free controller (RMFC) is designed to realize a compromise between the maximum-power-point trajectory tracking under realistic wind energy exploitation, the desired rotor speed reference trajectory tracking and the need of decreasing the solicitations induced in the mechanical structure of the turbine. The RMFC uses only the turbine output measurement and does not require knowledge of turbine parameters. It has a very simple structure and can be easily implemented in computer control systems. The effectiveness and robustness of the proposed control approach are proven by numerical simulations. Keywords: wind turbines, power regulation, recursive control algorithms. 1. INTRODUCTION Wind turbines function in extremely dynamical and unpredictable environmental conditions. The energy production is entirely dependent on the wind therefore an efficient extraction of the energy implies a good knowledge of the wind characteristics. The stochastic nature of the wind and the difficulties encountered by the scientists in conceiving predictions regarding wind’s speed, intensity or direction in time, make the energy production be sensitive to any sudden wind burst. The goals for the control of the wind turbines depend on the working regimes and will be briefly defined in the following sections. We have focused our interest on finding the command that would provide the optimum power level with the minimum mechanical stress sensed by the turbine’s physical structure within turbulent conditions and as well as in situations of steady winds. Many control solutions are proposed in the literature, all being conceived for a specific model, simplified or sophisticated, linear or nonlinear. In general, the existing control methods are based on the use of linear models supposed to closely approximate the nonlinear turbine dynamics [Lescher et al. 2006], [Yao et al. 2009], [Li et al. 2008], [Hand 1999]. Linear models allow obtaining relatively simple and analytical solutions (LQR, poles placement [Nourdine et al. 2010], [Munteanu et al. 2008], [Bianchi et al. 2007]), which are easily implementable in practical applications. However, the linear models bring the disadvantage that the controller’s performances are guaranteed only for the particular operating point specified in the linearization phase. On the other hand, the nonlinear model based methods can ensure larger operating regimes [Kebairi et al. 2009], [Boukhezzar and Siguerdidjane 2010] but their performances are strongly dependent on the availability of an accurate turbine model. In [Hand 1999] a comparative analysis of a PID controller designed for both linear and nonlinear models of a wind turbine is presented. In both cases, though, the controller is highly dependent on the model parameters. In this paper, we propose to utilize a Recursive Model Free Controller (RMFC) [Wang et al. 2010a], [Wang 2011] for horizontal variable speed wind turbine systems. The particularity of our controller is that it only uses the system output measurements and it does not require knowledge of the turbine parameters. Its structure is recursive, proportional (P) but with a time varying gain and it can be easily implemented in computer control systems. We shall demonstrate the performances of this controller for the wind turbine whose nonlinear model is briefly detailed in Section 2 of the paper. In Sections 3 and 4 the control objectives are formulated and a RMFC for a wind turbine system is developed. The controller performances are analyzed in Section 5, and at the end in Section 6, some concluding remarks are given.