Sistema de control auto-optimizable (SOC) para maximizar la extracción de potencia y minimizar las cargas en las aspas de un aerogenerador de pequeña escala

Abstract

Wind energy is the fastest growing technology in market sharing and efficiency. Wind energy conversion systems (WECS) transform kinetic energy from wind to electric energy through a combination of wind turbines (WT) and electrical generators. There are several possible configurations of wind turbines and electric generators, which allow different control approaches to regulate the energy output of WECS that mainly are: stall control, and pitch angle control. For any case, several approaches based on classic control and modern control have been successfully applied to the solely objective of improving the power capture of WECS, while, in most cases, leaving unattended other important operating variables, such as the blades’ mechanical stress. This research work proposes a methodology of design and testing of a self-optimizing control (SOC) algorithm to maximize the power output of a WECS while reducing the mechanical stress of the blades of the WT. This improved operation of the WECS is achieved by minimizing a multi objective cost function that relates the output power and the stress factor. The cost function calculation is performed by using a combined model of the blade element momentum (BEM) and thin wall beam (TWB) approaches. Since the SOC deployment implies low computational cost due to an optimization space reduction via a matrix projection applied to a measurement vector, the design of this algorithm requires off-line calculation of this projection matrix, H. SOC optimizes operation of the WECS in the presence of uncertainty or disturbance, which is assumed to be the variation of the wind speed, by controlling a carefully chosen combination of measurements to a constant set point. The performance of the WECS when operating with the SOC is compared with a baseline classic controller (BCS), and a predictive control algorithm (NMPC). Simulation results show the feasibility of the controller for real-time operation, as well an improved operation in comparison with the baseline controller.