Neural Predictive Controller for Hydraulic Power Transmission in Wind Turbine

Abstract

Fluid power transmission for wind turbines is quietly gaining more interest. The aerodynamic torque of the rotor blades is converted into a pressurized fluid flow by means of a positive displacement pump. At the other end of the fluid power circuit, the pressurized flow is converted back to torque and speed by a hydraulic motor. The goal of this paper is to develop a general dynamic model of a fluid power transmission for wind turbines, in order to gain better insight on the dynamic behavior and to explore the influence of the main design parameters. A fluid power transmission is modeled for a wind turbine with 1MW rated power capacity. This mathematical model can be used for simulation of the process using AUTOMATION STUDIO 5.2. Further the model has been approximated as a transfer function model using system identification toolbox available in MATLAB software. Neural network based predictive control (NPC) is applied to the mid-sized hydrostatic wind turbine model for maximizing power capture. The effectiveness of NPC is compared with PI controller.