Power quality improvement by dynamic voltage restorer and unified power quality conditioner using fuzzy logic

Abstract

Voltage sag and harmonics have significant negative impact on power quality. Propagation of distorted waveform in the system ultimately instigates a range of power quality issues. Protection against this situation is necessary because it adversely affects reliability and quality of power supply. Custom Power Devices (CPD) such as Dynamic Voltage Restorer (DVR) and Unified Power Quality Conditioner (UPQC) provide a level of reliability and power quality that is urgently required. The basic purpose of this thesis is to implement new control strategies to mitigate voltage sag/swell and to suppress harmonics to evaluate CPD performance. The CPD system is controlled by Fuzzy Logic Controllers (FLCs) with 49 rules and 25 rules, respectively and their performances are compared with Proportional Integral (PI) controller. This thesis focuses on FLC with fewer rules to avoid complexity, reduce computation time and consume less memory space. The proposed strategy depends on d-q transformation, phase-locked loop synchronization and constant DC link voltage. Simulation results depict better reliability of fuzzy logic techniques with non-linear loads when compared with PI technique. The capability of DVR and UPQC using FLC especially with fewer rules is demonstrated using test models built in Matlab/Simulink software. Both equipments potentially restore sags and improve overall harmonic profile while reducing total harmonic distortion to values within the prescribed criterion set by IEEE standards. The proposed FLC approach with reduced rule base, is superior to PI and FLC with 49 rules, assures effective performance, with normalization and tuning of parameters especially the membership functions and scaling factors. It yielded excellent voltage profile with best possible compensation during system contingencies, conforming to the IEEE standards that affirm FLC as an effective solution for power quality problems.