POWER QUALITY IMPROVEMENT IN GRID CONNECTED DISTRIBUTION SYSTEMS USING ARTIFICIAL INTELLIFENCE BASED CONTROL ALGORITHMS

Abstract

These days the Power Quality has become the key issue in electrical distribution power systems. The power quality is deteriorating because of the extensive utilization of non-linear loads and power electronic devices like solid-state switching devices which injects the current harmonics in the distribution systems. These harmonics leads to increase in power loss in transmission and distribution network. In this thesis performance analysis of artificial intelligence-based controllers such as Artificial neural network (ANN) and Adaptive Neuro Fuzzy Inference System (ANFIS) for shunt compensation has been studied for power factor improvement, current harmonics reduction and reactive power compensation. These control methodologies are executed in MATLAB Simulink 2018a. Their comparison analysis with the conventional time domain Instantaneous Reactive Power Theory (IRPT) has been analysed on the basis of various critical factors such as Total Harmonic Distortion Factor (THD), peak overshoot, settling time, convergence rate and the settling time of dc side capacitor voltage. The simulation results obtained demonstrates that ANFIS converges faster than ANN and IRPT under steady state and dynamic changing load conditions and has THD of around 1.06% obtained.