Augmented Complex Least Mean Square/Fourth Algorithm for Adaptive Frequency Estimation

Abstract

In order to solve frequency estimation of three-phase unbalanced power system, we propose an augmented complex least mean square/fourth (ACLMS/F) algorithm. The least mean square/fourth (LMS/F) algorithm combines advantages of least mean square (LMS) algorithm and least mean fourth (LMF) algorithm, which has superior performance in noisy environments. The Clarke's αβ transformation was used to get the information of three-phase voltage signal simultaneously, so the proposed algorithm will have better robustness. To improve the performance in unbalanced power systems, the widely linear modelling is introduced into the traditional complex-valued adaptive algorithm. The performance of ACLMS/F is texted in some situations usually appearing in the power system.