SOFT COMPUTING TECHNIQUES FOR CONTROLLING D-STATCOM IN DISTRIBUTION SYSTEM

Abstract

In recent years, there has been an exponential increase in energy demand in domestic, commercial and industrial sectors. Applications of electronically controlled devices provide automatic control in modern power system; yet due to their non linear characteristics, the current or voltage waveform deviates from the ideal sinusoidal waveform. Arc furnaces, converters, variable speed AC drives etc inject harmonics into the power system and affect the working of other sensitive devices. This may lead to partial or complete rupturing of equipment and blackouts. The presence of harmonics increases the transformer temperature, leading to equipment heating and increased losses. Sudden addition or removal of large load or capacitor bank results in voltage fluctuation causing sag and swell in system voltage. These problems are collectively grouped as power quality issues and the researchers are working worldwide to find solutions to these power quality problems for improving the efficiency and reliability of the power system. There are various devices like D-STATCOM, DVR and UPQC which are being used to eliminate different power quality problems. In this thesis, a three-phase, three leg D-STATCOM structure has been studied and analyzed. Conventional control theories viz. Synchronous Reference Frame Theory (SRFT) and Power Balance Theory (PBT) have been implemented in simulation model in MATLAB using SIMULINK and Sim Power System (SPS) toolboxes. Experimental verification of PBT has also been presented. The schemes have been modeled using Sim Power Systems and performance of D-STATCOM is studied for variety of loads under varying load conditions. The conventional control is replaced by Fuzzy logic controller (FLC) and simulation results are presented for the system with different set of rules. 49, 25 and 9 rule based FLC controllers have been designed and their performance studied for the same system over a wide variation in loading conditions. The performance of FLC is compared with PI controller. FLC requires high computational time therefore Adaptive Neuro-Fuzzy (ANFIS) controller is also designed and studied. Using suitable inputs, error and change in error over the DC link voltage, ANFIS controller is first trained. Once the performance of the ANFIS controller is satisfactory, the rule base developed by it is used for D-STATCOM control. The entire control scheme is simulated in MATLAB environment and also tested experimentally for a prototype system. The results of different control schemes such as conventional (SRFT and PBT), soft computing techniques like fuzzy and ANN have been used for D-STATCOM control in iii this thesis work. Power quality problems such as harmonic reduction, power factor correction, load balancing and voltage regulation have been addressed.