Filtering‐Based Key Term Separation for Nonuniform Sampling Hammerstein Systems With Colored Noises: Application to Wind Power Plants

This paper addresses the parameter identification problem for Hammerstein systems involving nonuniformly sampled input‐output data and colored noise. Filtering techniques are applied to preprocess raw input‐output data, aiming to mitigate the colored noise on identification accuracy. To address the parameter coupling issue caused by nonlinearity, the key term separation technique is employed, through the reconstruction of the system output, enables the model to be successfully transformed into a linear‐in‐parameters form. Incorporating the negative gradient search, a filtered key term separation gradient‐based iterative algorithm is derived. Finally, the numerical simulation results verify the effectiveness of the proposed algorithm in estimating system parameters. Meanwhile, the real‐world wind turbine case study ultimately validates the algorithm's practical effectiveness.