A New Approach for EPLL-Based Frequency Estimation under Severe Disturbances

The number of non-linear loads connected to power grids is experiencing a fast growing in the past decade as a consequence of alternative sources insertion, such as solar and wind generators. These changes have introduced multiple disturbances to the desired sinusoidal power signal, causing the estimation algorithms to lose accuracy or even fail. Classic and enhanced phase-locked loops, widely employed as frequency estimation methods, are sensitive to high amplitude noise and interharmonics, which has became an important issue for their application on this new power grid reality. This paper proposes an innovative method for improving the EPLL robustness against noise and interference, especially harmonic and interharmonic components, both static and time varying. It consists on a pre and post processing filter for EPLL frequency estimation based on Recursive Hann Window Discrete Fourier Transform - RHWDFT - which has been successfully applied as a high-selectivity filter implementation with low computational complexity, and an angular frequency smoother algorithm, which reduces the EPLL estimated angular frequency ripple. Several computer simulations concerning severely distorted input signals were done, and the results show that both interference rejection and frequency estimation accuracy have been improved. This methodology can also be extended to other time-domain frequency estimation applications.