Doubly fed induction generator identification based on Kalman filtering in the presence of spurious data

Abstract

This paper presents a methodology for parameter identification of a doubly fed induction generator (DFIG) in the presence of spurious data. DFIG is widely used in the electrical energy production using wind; one problem that the control system for these machines faces is the variability in the parameter values, and optimal performance for this control system is then hard to achieve. Besides, if the sensory system is not reliable, incurring in measurements with high uncertainty may be very common. To perform the parameter identification, three sequential Kalman filters are used, two of them are the dual Kalman filter, and another is the robust statistic Kalman filter. The methodology was implemented in Matlab, showing that the method is not affected by these data, obtaining residual errors smaller than 1.2% for the DFIG identification in the presence of these data.