Combined Feedback–Feedforward Control of Wind Turbines Using State-Constrained Model Predictive Control

摘要

An application of model predictive control (MPC) to the wind turbine collective pitch and torque control problem in full-load operation is presented. The applied controller is able to include upstream wind speed information, e.g., from light detection and ranging measurements. Furthermore, a method is presented to include constraints on the turbine states in the problem formulation, such as on rotor speed, even in the presence of unmeasured disturbances. In a simulation study, the effects of both preview control and state constraints are evaluated for three different scenarios: normal operation, gusts, and grid-loss load cases. It is found that preview control provides significant benefits in normal operation and under gust conditions. It is further found that including state constraints in the MPC formulation can be used to avoid unnecessary shutdowns due to the violation of the overspeed limit or to control the turbine effectively when the turbine needs to run at a reduced level of generator torque.