Harmonic current detection and suppression based on neural network

The distributed power generation systems based on the renewable energy sources, such as solar photovoltaic, wind power, geothermal power, have been the focus of the world in the latest years. The grid-connected inverter is important power electronic implements, used to transfer the power from the DPGS to the public grid, and the quality of its output power has been studied a lot. The repetitive control technique based on the internal model principle, can achieve low steady-state tracking error and total harmonic distortion (THD) in AC systems, so it is widely adopted in the inverter systems. However in the practical applications, there are still some issues needs solving when using the repetitive control technique. To suppress the current distortion, a control strategy of fundamental current and harmonics controlled separately is adopted. On the synchronous rotating coordinate, the fundamental component is direct value and the harmonic component is alternating value. Proportion-integral (PI) values above which can achieved output controller and proportion resonant (PR) controller are adopted to control the power adjustment and zero error control of harmonic restrain.