DISTRIBUTED MODEL PREDICTIVE CONTROL OF A WIND FARM WITH CLUSTERING

Abstract

This work gives a concise overview of the role that distributed model predictive control has el the development of the
advanced wind turbine control algorithms. The benefits of the model predictive control compared to conventional
controllers convoluted in wind turbine control are defined. Wind turbine model predictive active power controller based
on identified piecewise affine discrete-time state space wind turbine model is designed. The designed D-MPC controller
showed better performance. A wind farm with ten wind turbines was used as the test system. Research were attend and
evaluated, which include the operation of the wind farm with the D-MPC under low and high wind conditions, and the
dynamic achieved with a wind turbine out of service. With the fast gradient method, the convergence rate of the D-MPC
has been significantly improved, which decrease the iteration numbers. Appropriately, the communication burden is
reduced.