Search

Single‐phase transformerless inverters are widely installed in grid connected photovoltaic systems due to their outstanding advantages, namely, high efficiency, low cost and high power density. The lack of galvanic isolation leads to common‐mode voltage (or current) issues, which are solved by means of modifications to topologies, control schemes and modulation strategies. This study proposes a novel multilevel inverter for single‐phase transformerless photovoltaic systems. The topology is based on the neutral‐point‐clamped inverter, and is aimed to attenuate the leakage ground current. The latter is caused by the common‐mode voltage fluctuations affecting the stray capacitances formed between the photovoltaic cells and the grounded frame. On the one hand, the proposed multilevel inverter can synthesise a seven levels output voltage, which reduces the total harmonic distortion. On the other hand, the magnitude of the output voltage can reach up to 1.5 times the DC input voltage, which eliminates the need of a voltage‐boosting stage for grid interconnection. This study also presents three dedicated pulse‐width modulation strategies, all of them based on the level‐shifted multicarrier sinusoidal pulse‐width modulation scheme. Numerical simulations and experimental tests in a 1 kW setup have been performed to evaluate the performance of the proposed topology [ABSTRACT FROM AUTHOR]

The detection of islanding effect is one of the important issues for photovoltaic (PV) power system since islanding is dangerous to utility equipment and workers, and result in severe injuries and death. The novelty of this study is combining extension neural networks and Chaos synchronisation (CS) on islanding detection of grid connected PV systems based on non‐autonomous Chua's circuit. Combining CS and extension neural network type‐2 (ENN‐2), this research is to propose a novel detection method at detecting and distinguishing the occurrence of islanding effect based on non‐autonomous Chua's circuit. Simulation and experimental designs through powersim (PSIM) were applied to mimic PV power system to demonstrate the effectiveness of the proposed method. Results show that the accuracy of ENN‐2 achieves 98.4% on detecting the islanding effect for PV power system. [ABSTRACT FROM AUTHOR]