Shunt active power filtering with reference current generation based on dual second order generalized integrator and LMS algorithm

Abstract

For reducing the carbon emissions from utility sector, it is important to improve the efficiency of network and electrical loads. Current harmonics are responsible for the degradation of the system as well as load efficiency. The propagation of current harmonics into the grid can be prevented with shunt active power filters (SAPFs). Reference current generation (RCG) is critical for SAPF operation. This paper presents the control of SAPF with RCG based on least mean square (LMS) and dual second order generalized integrator (SOGI). The two SOGIs separately process grid voltage and load current to determine the respective fundamental in-phase component (FPC) and fundamental quadrature component (FQC). FPC and FQC of grid voltage are used to determine the phase angle of fundamental grid voltage. With LMS algorithm the distorted current can be decomposed into fundamental active (FA), fundamental reactive (FR) and harmonic components. However, the dynamic response of LMS based estimation is poor when the same learning rate is used for all components. FQC of load current, estimated with SOGI, can be processed by LMS algorithm to separate FA and FR components. RCG can be implemented with the computed phase angle and FA component. In LMS algorithm, as only two components are to be separated, two different learning rates can be easily specified. This results in faster dynamic response for LMS algorithm and has consequent positive impact on RCG and SAPF operation. Also, the computational complexity is considerably reduced. The performance of SAPF with the proposed extraction algorithm is analyzed for different operating conditions and comparative analysis is carried out.