Adaptive Filter to Compute the Reference Current on Shunt Active Power Filters

In this paper a new Adaptive Notch-Kahnan Filter (ANKF) is proposed to estimate harmonics and symmetrical components in distorted and unbalanced electrical signals. The new filter is an adaptive system that includes a Notch filter in cascade with a Kalman filter. The Kalman filter outputs are fed back to the Notch filter in a recursive process guaranteeing optimal signal parameters estimation and fast convergence without the use of a synchronization tool such as a phase-locked loop (PLL). The ANKF can be configured either for single-phase or three-phase applications and has two operation modes: normal mode and enhanced mode. The first one is a steady-state mode and guarantees the filter's ability to suppress the undesired signal components. The second one is a transient mode which accelerates the filter convergence every time that it detects a change in the steady-state signal, so that the new steady-state signal can be obtained faster. The filter feasibility has been successfully tested on different waveforms, obtaining a good and fast response. In addition, the proposed filter has been used in an experimental setup to compute the current reference in a shunt active power filter. Copyright (C) 2011 Praise Worthy Prize S.r.l. - All rights reserved.