Disturbance Rejection and Uncertainty Analysis in Wind Turbines Using Model Predictive Control

For effective wind turbine operations, it is essential to maintain the power limit and reduce the stress on the drive train in the presence of disturbance and uncertain conditions. In our work, we propose a Model Predictive Control (MPC) framework with quadratic cost functions, incorporating control input and state constraints to mitigate the challenge of disturbance rejection and uncertainty analysis for the wind turbine operation. We have tailored the algorithm to the practical parameters of the National Renewable Energy Laboratory's (NREL) Controls Advanced Research Turbine (CART) model. We illustrate the impact of wind disturbances on achieving the optimal control law and evaluate the performance of integral MPC in disturbance rejection for the wind turbine operation, comparing it with the constrained optimal control law outcomes. The simulation results also show the efficacy of integral MPC for the uncertainty in the initial conditions of the wind turbines. This is shown by the propagation of the first two moments, i.e., mean and variance, for the states of the wind turbine. Further, we obtained the control law and mean-variance propagation for the variation in disturbance intensity. The overall results prove the efficacy of using the MPC framework for uncertainty analysis and disturbance rejection to obtain optimal operation in wind turbines.