A Variable Step Size, Adaptive Algorithm Applied To Active Power Filter

A number of conventional time-varying step size algorithms have been tried on the active power filter (APF), and experiments indicate that their performance is highly sensitive to the noise disturbance. The paper presents a novel adaptive algorithm using the weighted average estimation of error signal e (n) to control the updating of the step size. It can provide a faster convergence at early stage of adaptation and ensure a small final maladjustment even under low signal-to-noise ratio (SNR) conditions. In this way we can make more flexible control between maladjustment and convergence time without the need to compromise one for the other. Theoretical analysis is also verified by simulation and experiment results