Partially Decoupled Adaptive Filter Based Multifunctional Three-Phase GPV System

This paper deals with partially decoupled adaptive Volterra filter (PDAVF) based control for a three-phase two-stage grid-interfaced photovoltaic (GPV) system. Besides maximum power extraction, the proposed control is having the potential of grid currents balancing, harmonics currents mitigation, reactive power compensation, and adaptive adjustment of DC bus voltage. The control technique is efficient in extraction of weight component of reference grid currents and an adaptation routine of filter weights uses the principle of fifth-order PDAVF with single-element observation vector implemented by using the method of least mean squares (LMS). An increased order of partial filter assures accurate estimation of filter weights by adaptive weight update at each filter step. The convergence is ensured by providing bounds on algorithm's step size. The algorithm overcomes drawbacks of Adaline-based LMS and LMF (least mean fourth) based weight estimations without imposing high computational burden. The switching losses in voltage source converter are minimized using adaptive DC bus voltage. Effectiveness of PDAVF is presented through simulation and test results. These results are found satisfactory with improved steady state and dynamic performances and total harmonic distortion of grid currents meet an IEEE-519 standard.