Neuroadaptive Power Tracking Control of Wind Farms under Uncertain Power Demands

Wind farm contains a large number of wind turbines, each of which is required to deliver certain amount of power so that the combined power from the wind farm is able to meet the total power demand. For such typical power tracking control problem, it is quite challenging to develop a computationally inexpensive and structurally simple solution. The problem is further complicated if the demanded power is unknown a priori and there exist modeling uncertainties as well as external disturbances in the system. In this paper, a neuroadaptive feedback control is presented. The barrier Lyapunov function based design technique is utilized to guarantee that the neural network (NN) training inputs are confined within a compact set such that the NN unit can maintain its learning/approximating functionality during the entire process of system operation. To address the issue of unknown power trajectory, an analytical model is proposed to reconstruct the unknown desired power profile. Both theoretical analysis and numerical simulation validate the effectiveness of the proposed method.