Complex-Valued Minimum Total Error Entropy Adaptive Algorithm With Fiducial Point for Power System Frequency Estimation

Recently, complex-valued adaptive filter has been utilized for frequency estimation of power systems. The augmented complex-valued minimum error entropy with fiducial points (ACMEEF) adaptive algorithm has achieved good estimation for power system frequency. However, it only considers the clean input signals. For noisy inputs, the ACMEEF algorithm will yield a biased estimation. To solve this problem, the augmented complex-valued minimum total error entropy with fiducial points (ACMTEEF) algorithm is proposed in this brief, which incorporates the MEEF criterion into the framework of total least squares (TLS) and performs well for noisy input and impulsive noise disturbance. In addition, the proposed ACMTEEF algorithm can still estimate the system frequency accurately under three-phase unbalance. Then, the theoretical performance of the ACMTEEF algorithm is analyzed, and the superiority of the ACMTEEF algorithm is demonstrated through simulation experiments.