Design of closed loop control for a wind turbine system coupled to a CV transmission system

Grid integration of renewable energy sources has proven to be a popular and challenging problem that has been extensively studied and continues to be a focus of interest. Most modern wind energy conversion systems utilize power electronics converters/inverters to maintain voltage phase, frequency and magnitude at the grid-dictated values. While power electronics is an expanding area of interest, currently available solutions report high failure rates and elevated monetary costs. In this paper we investigate the dynamic analysis of a gearless wind turbine coupled to an induction generator based on the dynamical understanding proposed by Ericson and Srivastava that describes both the steady state and the shifting behaviours of the V belt CVT. The Model uses the dynamics published by Carbone-Mangialardi to explain this relationship during creep mode shifting, and the dynamics by Shafai in Slip Mode shift, which provides thorough detail on the inertial interactions between the belt and the pulleys of CVT. Using Matlab/Simulink, we incorporate the CVT model into a wind turbine model coupled to an induction generator. The entire turbine/rotor - CVT/generator is coupled to the grid through the conventional grid- and rotor- side converters. By controlling the driver axial forces of the CVT we intend to be able to control low- and high- speed shaft speeds (i.e. perform speed control) to obtain maximum wind energy capture before the wind cut-out speed and provide an alternative to pitch control afterward. The intent is to understand how control inputs of the CVT affect power through the entire drivetrain to meet the objectives of: a) Maximal power extraction from the wind b) Better quality power for grid integration c) Tracking the grid demands without degrading the CVT performance The results for the overall integrated powertrain are presented and discussed in detail with the CVT and the induction generator operated in closed loop configuration. The simulations were all performed utilizing real wind data taken from a met tower in the South plains area, the data was sampled at 50 Hz.