Artificial Neural Network based Solar Energy Integrated Unified Power Quality Conditioner

Abstract

ABSTRACT The utility grid is prone to power quality issues due to the advent of power electronic devices and integration of Distributed Energy Resources. The power quality has to be maintained despite variation in load under steady and fault conditions. Unified Power Quality Conditioner is a FACTS device that improves the power quality at the Distribution side of the utility grid. In Unified Power Quality Conditioner, reactive and real power compensation is carried out simultaneously improving the power quality. The proposed control algorithm for Unified Power Quality Conditioner is a hybrid combination of Reactive Power Control and Unit Vector Template. The Unified Power Quality Conditioner is integrated with Distributed Energy Resources like solar energy to minimize the power rating of converters and meet the power demand. The reinforced learning algorithms have been effectively increasing the performance of the power electronic devices; among them, the most commonly used is the Neural-Network algorithm. The Artificial Neural Network controller for the solar integrated Unified Power Quality Conditioner system improves the power quality when compared to the conventional controllers by self-adapting themselves to the environmental needs. The system is tested under both balanced and unbalanced load conditions with MATLAB-SIMULINK. The per-unit system has been used for the analysis purpose to minimize the complexity. The hybrid control of the series and shunt converters supported by solar integration at the DC link proves to eliminate distortion caused by the non-linear load. The system is subjected to a momentary voltage sag/swell as per IEEE 1159 standards. The momentary sag/swell that varies between 0.5 seconds to 3 seconds has been made to ride through less than 50 milliseconds as per the Computer and Business Equipment Manufacture Association curve. The load side current harmonics have been minimized inevitably by the proposed control methodology to ensure reliability and good power factor at the Distribution side. A comparative analysis between the conventional Proportional-Integral controller and Artificial Neural Network controller in solar integrated Unified Power Quality Conditioner proved the Artificial Neural Network controller improved the harmonic reduction of various orders by an average of 12.72%. The Total Harmonic Distortion minimization for various harmonic orders was brought under the limits of IEEE 519 standards with Artificial Neural Network controller.