Hybrid Conjugate Gradient Backpropagation of GCPV based DSTATCOM for Power Conditioning

Abstract

This paper studies the performance of a hybrid conjugate gradient backpropagation (HCGBP) grid-connected solar photovoltaic (GCPV) based DSTATCOM. This paper proposes a hybrid control algorithm of instantaneous reactive power theory and conjugate gradient backpropagation neural network for an application of a grid-connected solar PV (GCPV) based DSTATCOM for three-phase three-wire system. The fundamental weighted value of active power components of load currents, which is necessary for estimating reference source currents, is extracted using a conjugate gradient backpropagation control algorithm. The performance of the proposed control algorithm has reduced the THD of the line current up to 1.32%. It is proven that HCGBP has better efficiency, faster response and easy to implement. The steady-state performance of the three-phase GCPV-DSTATCOM under non-linear load has been analysed through simulation and Hardware-in-loop (HIL) simulation based on real time DSP system using Texas Instrument TI C2000 32-bit microcontroller in MATLAB/Simulink. Furthermore, the simulation results have shown that the THD of the line current at the PCC has reduced less than 8%, according to the IEEE standard 519:2014.