Reweighted L1 Norm Penalized LMS Fourth Algorithm of Solar Grid Interfaced System for Alleviating Power Quality Problems

摘要

This study investigates an applicability of a reweighted l₁ norm penalized least mean square fourth (RL₁-LMSF) algorithm and an improved multivariable filter based frequency locked loop (IMVF-FLL) through the performance demonstration of a three-phase solar-grid interfaced system in a double stage topology. The effectiveness of the system is ascertained by considering grid voltages unbalance, voltage sag/swell, unbalanced dc offset, unbalance in loads, and variation in solar irradiance. This RL₁-LMSF algorithm adaptively extracts fundamental component of nonlinear load currents for estimation of reference grid currents with fast convergence and least steady-state error. An IMVF-FLL is used to elicit positive sequence components of grid voltages at abnormal grid conditions. The adaptability to input frequency deviations, harmonics suppression, dc offset rejection, and fast response of IMVF-FLL, during dynamic conditions, enables accurate estimation of unit templates. These unit templates maintain sinusoidal and balanced reference grid currents at varying conditions. The harmonic distortions in grid currents are strictly following the standard of IEEE-519. The photovoltaic array is kept at maximum power point using a boost converter together with a technique based on incremental conductance. The system is simulated in MATLAB environment, and test results recorded using a hardware setup emphasize its relevance.