Full state feedback digital control of WECS with state estimation by stochastic modeling design

Abstract

With increased structural flexibilities of modern wind energy conversion systems (WECSs), drive-train torsional oscillations due to high wind disturbances can be hazardous to power systems. It is particularly important to design control schemes that properly account for the flexible modes of the turbine, and maintain stable closed-loop behavior of the WECS. To capture the essential features of the problem for controller design, the approach in this paper entails modeling the various WECS subsystems: the wind speed as a stochastic process generated by an autoregressive moving average (ARMA) model and the drive-train as a multimass system with an elastic shaft linking the turbine with the asynchronous induction generator (IG). The proposed control paradigm is based on the linear quadratic Gaussian (LQG) to damp these undesired oscillations, and utilizes only one measurable feedback signal: generator speed deviation. An optimal state estimator provides a good estimation for the aerodynamic torque thereby determining the target trajectory to which the controlled system should converge. The objective is to optimize aerodynamic efficiency in low to medium winds, and add damping to the drive train at above-rated wind speeds. A pitch controller ensures the maximum power constraint is respected by preventing rotor overspeed.Index Terms-Aerodynamic power, ARMA wind model, drive train torsional torque, LQG, pitch regulation, state estimation.

Optimal LQG Controller for Variable Speed Wind Turbine Based on Genetic Algorithms

“…That type of control shows useful properties of good performance and robustness in controller design applied to wind energy converters system [1]. Although the application of LQG control for DFIG wind turbines is not new, recent research report using LQG controllers successfully [2,3,4,5].…”