Intelligent Controlled Dynamic Voltage Restorer for Improving Transient Voltage Quality

A dynamic voltage restorer (DVR) is developed to stabilize the three-phase load voltage under the sudden grid voltage distortion conditions, including the voltage sag, swell and unbalance. The DVR is composed of a DC
$\backslash $
AC inverter and a DC-link capacitor regarded as the energy storage element. When the sudden grid voltage distortion occurs, the instant power to compensate or absorb is realized by the DC-link capacitor of the DVR resulted in the degenerate stabilization performance and the poor transient voltage quality of the load. Hence, the DC-link voltage control of the DVR for maintaining the constant DC voltage plays an important role in stabilizing the three-phase load voltage. In this study, to improve the transient responses of the three-phase load voltage and the DC-link voltage under the sudden grid voltage distortion conditions, a novel recurrent compensation petri fuzzy neural network (RCPFNN) controller is firstly proposed to replace the conventional proportional-integral (PI) and fuzzy neural network (FNN) controllers for the DC-link voltage control of the DVR. The detailed network structure and the online learning algorithm of the proposed RCPFNN controller are derived. Moreover, some experimental results are given to verify the feasibility and effectiveness of the DVR using the proposed RCPFNN controller for improving the transient voltage quality of the load.