Adaptive phasor estimators based on recursive least-squares

Recursive algorithms, such as Recursive Least Square (RLS), have a very fast initial convergence rate, but the algorithm gain becomes too small just after a few iterations. This reduces the algorithm´s adaptability. This paper summarizes the main techniques to overcome this drawback which are based on the manipulation of the covariance matrix. Moreover, it is introduced an alternative adaptive technique designated Modified Random Walking (MRW) suitable for digital relaying applications. This provides the algorithm´s capability to track amplitude and phase of currents and voltages collected from power grids under faulty conditions. The performance of the algorithms is evaluated using computer simulations.