Subsynchronous Resonance in Doubly-Fed Induction Generator Based Wind Farms

Licentiate thesis, 2016

The objective of this thesis is to investigate the risk for instabilities due to SubSynchronous Resonances (SSR) conditions in large wind farms connected to series-compensated transmission lines. In particular, the focus is on Doubly-Fed Induction Generator (DFIG) based wind farms. Analytical models of the system under investigation are derived in order to understand the root causes that can lead to instabilities. A frequency dependent approach, based on the Nyquist criterion, has been applied in order to investigate the risk for SSR in DFIG based wind turbines. Through this approach, it is shown that the observed phenomenon is mainly due to an energy exchange between the power converter of the turbine and the series compensated grid. This phenomenon, here referred to as SubSynchronous Controller Interaction (SSCI), is driven by the control system of the turbine, which presents a non-passive behavior in the subsynchronous frequency range. The different factors that impact the frequency characteristic of the wind turbine, thereby making the system prone to SSCI interaction, have been investigated. Through this analysis, it is shown that in a DFIG wind turbine, the current controller in the rotor-side converter plays a major role and that the risk for SSR increases when increasing its closed-loop bandwidth. In addition, it is shown that the output power generated from the wind turbine has an impact on the frequency characteristic of the turbine. Time-domain studies are performed on an aggregated wind turbine model connected to a series compensated transmission line with the objective of verifying the analytical results obtained through frequency-domain analysis. Based on the theoretical analysis, mitigation strategies are proposed in order to shape the impedance behavior of the wind turbine in the incident of SSCI. The effectiveness of the proposed mitigation strategies are evaluated both theoretically through frequency domain analysis and using detailed time-domain simulations.