Dynamic voltage restoration using neural networks for grid-connected wind turbine

Abstrak

Wind energy is being integrated into the grid as a renewable energy source to meet the world's electricity needs. Grid-connected wind turbines are often disrupted by grid fault problems. Fault ride-through (FRT) ability has become the most important grid connection necessity for wind energy conversion systems (WECS). In the event of a voltage dip fault, the low voltage ride-through (LVRT) capacity is an imperative key to successful grid integration. This paper proposes a dynamic voltage restorer (DVR) controlled through an artificial neural network (ANN) to improve the LVRT capability of a grid-connected wind turbine (WT) based permanent magnet synchronous generator (PMSG). The DVR injects series voltage into the system through a series-connected transformer. The DVR can then restore the voltage to the pre-fault value. The injection transformer is connected to the line linking the PMSG-based wind turbine output to the utility grid. Design and simulation of the low voltage ride-through applied to symmetrical and asymmetrical fault conditions were performed in MATLAB/Simulink software. Simulation results approve that the performance of the technique fully demonstrates its effectiveness and practicality.