Wind farm active and reactive power management

Abstract

ABSTRACT One of the most critical aspects in the wind farm sector is a power management throughout different wind generators. This paper proposes an improvement power management strategy based on the hybridisation of a Proportional Integral controller with aerodynamic power prediction using an Artificial Neural Network (ANN). The motive of this study is to create a simple and robust algorithm that defines short-term wind power prediction. A set of recent wind speed measurements, obtained between 1995 and 2004 by the Algerian south station (ADRAR), are used to train and test the data set. The suggested algorithm has been successfully implemented for the power management of the wind farm using Matlab Simulink. The achieved results validate the proficiency of the modified algorithm which gives a good solution for the saturation problem in the wind generator level. Abbreviations: ANN, Artificial Neural Network; AMO, Algerian Meteorological Office; MAPE, The Mean Absolute Percentage Error; DFIG, doubly fed induction generator; TSO, transmission system operator; PI, Proportional Integral; PCC, Point of Common Coupling; MPPT, Maximum Power Point Tracking; P&Q, Generator Active and reactive power diagram