A globally convergent wind speed estimator for windmill systems

Abstract

An estimator of the wind speed of a wind turbine coupled to a generator is proposed in this paper. Wind speed enters into the generator dynamics through a highly nonlinear function, hence we are confronted with a difficult problem of estimation of a nonlinearly parameterized system. To solve this problem we use the technique of immersion and invariance, recently introduced in the literature. It is assumed that the rotor speed and electrical torque of the generator are measured, which is the case for the machines typically used in this application. The result is of interest for the design of controllers of maximum power extraction, where the knowledge of the wind speed is necessary to express the control objective as a speed tracking problem. Detailed computer simulations are presented to assess the performances of the proposed estimator and a certainty equivalent proportional plus integral controller.

Maximum power point tracking algorithms for wind power generation system: Review, comparison and analysis

“…However, the signals of rotational speed and rotor position estimated by the sliding mode observer would appear chattering. Soltani et al 27 denied the method of measuring wind speed in the engine room and evaluated the dedicated estimators for estimating wind speed given by many algorithms, which include the power balance estimator, 28 Kalman filter (KF) estimator, 29 extended Kalman filter (EKF) estimator, 30 disturbance accommodating control (DAC) estimator, 31 unknown input observer (UIO) estimator, 32,33 and immersion and invariance (I&I) estimator 34 . Mérida et al 35 proposed an equivalent wind speed estimation method based on numerical calculation such as the Newton–Raphson method after calculating the aerodynamic torque. …”

Active Power Control of Waked Wind Farms: Compensation of Turbine Saturation and Thrust Force Balance

“…The demanded power from the TSO is set to reach the maximum of 70% of the total rated power. The available power herein is extrapolated using the maximum power coefficient C p max and illustrated by P avail i (v r ) = min(0.5ρπR 2 v 3 r C p max , P rated i ), where the effective wind speed v r is estimated by the I&I technique [29]. To assess the performance of the proposed controller, the setting case 1 with only the CCL on and the setting case 2 with both the CCL and TCL on are simulated at different average inflow wind speeds.…”

Model-free estimation of available power using deep learning

“…Similar methods are also reported in Henriksen et al (2012), where the dynamic inflow model is included. Besides the Kalman-filter-based approaches, some studies (Ortega et al, 2011(Ortega et al, , 2013) used a more advanced state estimation technique of immersion and invariance to construct a wind speed estimation with proof of global convergence under certain assumptions. For more details and further information on wind speed estimation, see Soltani et al (2013) and references therein.…”