Voltage quality improvement in electrical distribution networks using dynamic voltage restorers: design, simulation and experimental tests of a robust controller

This paper presents a simulation and implementation of a proportional integral (PI) and adaptive neuro-fuzzy inference system (ANFIS) control scheme of single-phase dynamic voltage restorer (DVR) for the mitigation of load voltage sag, swell and harmonics. The objective of the control strategies is to regulate the voltage of the DVR via an injection transformer to compensate for the voltage required and maintain the load voltage at a constant value. However, the conventional compensation techniques are grouped only on the load voltage and the DVR itself. The aim is to regulate the injection voltage of the DVR to compensate the grid voltage via the injection transformer in addition to maintaining the load voltage stable. These methods have been adapted to sinusoidal references and is resistant to sags, swells and harmonics. The proposed control strategies of the DVR are initially evaluated in simulations under MATLAB/Simulink and then validated on a laboratorial prototype of the single-phase DVR. n in-depth analysis across different controllers would show the performance and their robustness in mitigating network power quality issues. The robustness of the proposed ANFIS controller is examined by comparative studies with the PI controller. We notice that the proposed ANFIS controller was able to obtain a lower total harmonic distortion than that of the PI method.