Parameter estimation of a doubly-fed induction generator in a wind generation topology

• Published in: 2012 47th International Universities Power Engineering Conference (UPEC) pp. 1 - 6
• Main Authors: Bekker, J. C., Vermeulen, H. J.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2012
• Subjects: Angular velocity; Generators; Mathematical model; Matlab; Modelling; Parameter estimation; Rotors; Simulink; Stator windings; Wind Turbine System
• ISBN: 1467328545; 9781467328548
• DOI: 10.1109/UPEC.2012.6398607

The installed capacity of wind generation systems has increased dramatically in recent years, thereby increasing the need for efficient model implementations for operational applications such as simulating system behaviour and grid interactions, identifying and quantifying ageing effects and performing condition monitoring. While the model topologies for conventional systems are well established, the methodologies for determining parameter values from online measurements require further research, especially for applications such as condition monitoring. This paper presents the results of an exploratory investigation to estimate the model parameters of a double-fed induction generator used in a classical wind generation system that includes a wind turbine blade, gearbox and simplified grid model. The investigation is conducted using dedicated Matlab implementations of C-code S-function models of the various system components, compiled as a Simulink library. This approach ensures highly efficient model implementations with fast simulation times as is typically required by parameter estimation processes. Results are presented for case studies performed on both ABC and DQ generator model implementations, with the generator either operated in isolation or as part of a wind generation system. The investigation considers two perturbation signals, namely step perturbations applied to the rotor angular velocity and stator voltages respectively. The results show that the electrical parameters of the generator can be estimated with good accuracy using these model implementations, operating topologies and perturbation signals.