A distribution grid tied multifunctional SPV system operating with control approach based on decoupled adaptive neural network

Abstract

This paper presents a single-stage, three-phase four-wire grid tied SPV (Solar Photovoltaic) energy conversion system with a decoupled adaptive neural network based control approach for power quality improvement at AC mains. The proposed solar PV energy conversion system, along with feeding fundamental active power to the grid, is capable of compensating linear and nonlinear loads with MPPT (Maximum Power Point Tracking), supply currents balancing, harmonics elimination, PFC (Power Factor Correction) and mitigation of neutral current. The proposed control approach is based on a decoupled adaptive neural network for each phase, which is used along with PV power feed-forward compensation. Decoupled adaptive neural network for each phase improves the convergence speed and provides fast dynamic response to sudden and severe load changes. The proposed grid tied SPV energy conversion system is implemented using DSP (Digital Signal Processor) and its detailed performance analysis is demonstrated.