Abstract--- The brushless doubly-fed induction generator (BDFIG) has no need for brushes and slip rings, low current harmonics and reduced maintenance cost, making it highly suitable for offshore wind power generation applications. Traditional control methods for BDFIGs include vector control and model predictive control. However, when machine parameters in the controller are affected by temperature, saturation, and other factors, the control performance deteriorates to varying degrees. To address this issue, this paper introduces an adaptive robust predictive current control strategy for BDFIGs based on an online-calculated ultra-local model (ULM), which eliminates the need for any motor parameters. The proposed method uses an ultra-local model to estimate the input voltage gain and total disturbances term through the sampled voltage and current in real-time. It then integrates this with deadbeat principle to achieve fast and precise regulation of the control winding current. Experimental results show that, compared to existing methods, this strategy offers excellent steady-state and dynamic performance when parameters are accurate and maintains good control even when parameters change, indicating strong parameter robustness.

robust predictive control,ultra-local model,brushless double-fed induction generator