Model-based condition monitoring of a doubly-fed induction generator slip ring component

Over the past decade the wind energy sector has grown worldwide from an installed capacity of 31.1 GW in 2002 to 273 GW in 2012. South Africa is forecasted to be one of the largest future developing markets for wind generation, particularly due to the need to introduce clean renewable energy into the generation capacity that currently consists predominantly of fossil fuel plants. This paper presents results of on-going research that focuses specifically on the condition monitoring of the slip ring component found in a doubly-fed induction generator, the most widely used generator topology throughout the wind energy sector. The slip ring component is maintenance intensive and that results in frequent unplanned outages and serious plant failures. Online condition monitoring of the slip ring component, especially in quantifying and trending ageing effects, has the potential to result in improved maintenance strategies and the prevention of failures. The approach investigated in this research involves the use of parameter estimation techniques to determine condition-related model parameters using C-code and S-function models implemented in Matlab and compiled as a Simulink library. The study aims to determine to what extent the parameters of the various subsystems can be estimated accurately enough from measurements of the terminal voltage and current waveforms to facilitate an online condition monitoring strategy. This paper presents results for one aspect of this study, namely condition monitoring of the slip ring component. Case studies are done on various theoretical- and laboratory data sets and the results of which are then compared and examined.